Political Economy of Islamic Banking Growth

Durham Research Online

Deposited in DRO: 17 July 2020 Version of attached le: Published Version Peer-review status of attached le: Peer-reviewed Citation for published item: Asutay, M. and Sidek, N.Z.M. (2021) 'Political economy of Islamic banking growth: does political regime and institutions, governance and political risks matter?', International journal of nance and economics., 26 (3). pp. 4226-4261. Further information on publisher's website: https://doi.org/10.1002/ijfe.2011

Publisher's copyright statement:

c 2020 The Authors. International Journal of Finance Economics published by John Wiley Sons Ltd. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Use policy

The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that: • a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.

Please consult the full DRO policy for further details.

Durham University Library, Stockton Road, Durham DH1 3LY, United Kingdom Tel : +44 (0)191 334 3042 | Fax : +44 (0)191 334 2971 https://dro.dur.ac.uk Received: 3 February 2019 Revised: 29 March 2020 Accepted: 18 June 2020 DOI: 10.1002/ijfe.2011

RESEARCH ARTICLE

Political economy of Islamic banking growth: Does political regime and institutions, governance and political risks matter?

Mehmet Asutay1 | Noor Zahirah Mohd Sidek2

1Durham Centre for Islamic Economics and Finance, Durham University Business Abstract School, Durham University, Durham, UK This article explores the political economy of Islamic banking by examining 2Economics Department, Universiti the impact of political regime types, institutional environment, government Teknologi MARA Kedah, Kedah, and political risk on the development of Islamic banking proxied by financing Malaysia or loan growth in the case of 16 Muslims majority countries with autocratic Correspondence and democratic regimes over the period of 2000–2013. The performance of Mehmet Asutay, Durham University Islamic banking loan growth is examined from three different perspectives— Business School, Durham University, Mill Hill Lane, Durham DH1 3LB, UK. political regime and institutions, governance and political risks in both regime Email: [email protected] settings. Results suggest that loan growth is positive and significant in democratic regimes where political and civil rights are predominant. In addition, loan growths are slower during election years but higher throughout pre-election year in democratic regimes, suggesting the opportunistic behaviour of incumbent government artificially boosting the economy in preparation for the upcoming election. It is also found that the good quality of public services, policy formulation and implementation, and credibility of government's commitment to realizing the policies are vital in ensuring positive loan growth. The lack of differences in the reaction of loan growth to the governance in democratic and autocratic regimes shows some convergence in the regimes despite having significantly different political perspective. This implies that political commitment of the ruling government is vital to set forth a strong and robust Islamic banking standing regardless of its political standpoint.

KEYWORDS electoral politics, governance, institutions, Islamic banking, loan growth, political regime and political risk

1 | INTRODUCTION Islamic law or Shari'ah compliant sphere to capture the demand for this growing market. Total Shari'ah compli- The unprecedented growth of Islamic banking and ant assets in the Islamic finance industry was reported as finance in recent years has prompted financial services $2.5 trillion in 2018 (Thomson Reuters, 2019) compared providers to expand their product portfolios towards to $861 billion in 2008 (Ernst & Young, 2009). The

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2020 The Authors. International Journal of Finance & Economics published by John Wiley & Sons Ltd.

Int J Fin Econ. 2020;1–36. wileyonlinelibrary.com/journal/ijfe 1 2 ASUTAY AND MOHD SIDEK statistics show that 70% the reported assets are with the spheres in the development of banking sector in general Islamic banking sector (Thomson Reuters, 2019) indicat- and Islamic banking in particular. ing overwhelming importance of Islamic banking within To initiate a broader debate on Islamic banking per- the existing shari'ah compliant assets. Over the past four formance by probing into the role of political economy decades, Islamic banking has thriven in Muslim majority and governance in fostering the development of Islamic countries and currently advancing its footing in non- banking as proxied with loan growth, this papers is con- Muslim majority countries worldwide. As of 2018, structed on the argument that performance and growth approximately 44.9% of global assets in Islamic banks are in banking sector is an endogenous outcome generated in the GCC countries, while Malaysia's shares in global by the type of regime, the quality of political institutions, Islamic banking assets is 10.8% (IFSB, 2019). political accountability and governance as part of public The aftermath of repeated financial crises over the policy apart from the usual micro and macro determining last two decades has unlocked the potential of Islamic variables. This leads to the hypothesis that better institu- banking as an alternative financing method and as a new tions lead to business-friendly policies, less red-tape, financial breed for the overall financial architecture. investor-friendly and less overall inefficiency, ultimately With the growing number of world Muslim population, enhancing financial development in the country gener- new Islamic banking and finance alternative is consid- ally. It is important also to note that since Islamic bank- ered having essential growth potential. Even in a consti- ing products mirror their conventional counterparts, it is tutionally secular state such as Turkey has in recent years hypothesized that their reactions towards institutional intensified its effort in championing Islamic banking and political variables would be the similar but relatively with plans to boost Islamic finance total assets to $100 small size of Islamic banking as a whole may affect their billion by 2023 (Reuters, March 15, 2015). The Turkish reactions towards political variables. government issued a $1.5 billion sovereign sukuk in 2012 This article is, therefore, developed around the follow- and licenced two state banks (Vakif Katilim and Ziraat ing questions: If the political economy of institutions mat- Katilim) to boost Islamic finance in Turkey in 2015, while ter, does loan (financing) growth behaviour, as an Turkey's share in the global Islamic banking assets indicator of Islamic banking performance, differ in the dif- remains at 2.9% only. In Kuwait, loans growth in Islamic ferent political regime and political institutions? To explore banking stands at 11.2% ($47.8 billion) compared to con- and examine this research question, this article compares ventional banking with only 7.5% between January and the behaviour of loan growth of Islamic banks in both September 2014 (Arab Times, 2015). According to the democratic and autocratic regimes, and, in comparison, World Islamic Banking Competitiveness Report 2016 with the conventional banks. The performance of Islamic (Ernst & Young, 2016), in 2014, Shari'ah compliant loans banks measured in terms of financing or loan growth is and deposits account for almost 54% in Saudi Arabia, examined from three different political economy frames— 33% in Bahrain, 54% in Kuwait, 25% in Qatar and exceed polity, governance and political risks in both regime types. $100 billion in Malaysia and UAE. Financing or loan growth is selected because it indicates This article extends the growing literature on Islamic growing businesses as a result of favourable business envi- finance from the political economy and institutional per- ronment propelled by proper institutional management. spective. While there is an extensive empirical literature A sample of 16 majority Muslims countries with auto- and practitioners' paper on Islamic banking, the impact cratic and democratic regimes is selected to examine this of institutions and political economy on Islamic banking case over the period 2000–2013. The sample of Muslim has relatively been underexplored. This general dearth of dominated countries led by democracies includes Turkey, academic work on the impact of institutions on Islamic Malaysia, Indonesia, Pakistan and Bangladesh, while banking stands in contrast with the increasing impor- autocratic regime is represented by the GCC countries, tance that Islamic banking has in Muslim majority as Sudan, Egypt, Yemen, Iran and Jordan. well as in non-Muslim majority countries. This article, Although the autocratic regime-based countries have hence, aims to explore and examine the impact of politi- greater Islamic banking total assets as can be seen in cal economy and institutional environment on the perfor- Figure 1, the data suggest that loan growth is higher in mance of banking generally and in Islamic banking, democratic regimes compared to autocratic regimes. specifically proxied by financing, namely loans.1 By com- Overall, we find that regardless of political regime type, paring conventional banking and Islamic banking in the reaction of loan growth towards political variables Muslim majority countries with a different ontological does not significantly differ, which could be possible due base and political philosophy with its relevant settings, to the impact of international financial integration which we aim to identify the role of political regimes and insti- eliminates these differences. The findings also suggest tutions, governance and political risk in different political that both institutionalization and competitiveness of ASUTAY AND MOHD SIDEK 3

FIGURE 1 Growth of Autocratic countries shari'ah compliant assets USD Yemen (billion). Source: The Banker Egypt (2013, 2014) [Colour figure can Sudan be viewed at Qatar wileyonlinelibrary.com] UAE Kuwait Bahrain Saudi Arabia 2014 Iran 2013 2010 2007

Democratic countries Bangladesh Pakistan Malaysia Indonesia Turkey 0 100 200 300 400 500 600 executive selection play essential roles since the frame- sample and two sub-samples indicating that durability of a work set by them would affect loan growth. However, fre- particular party or monarchy may impede the adoption of quent changes in executive may result in policy change new technologies and ideas especially in banking, which, which is not an ideal situation in promoting loan growth. in turn, slows down loan growth. Finally, results depict that Therefore, all election dummy variables are found to political risks variables impacted loan growth differently have a negative impact on loan growth. In the Islamic across the samples. In the overall sample, government sta- banks in democratic countries sub-sample, the results bility, internal conflict, law and order and democratic demonstrate that loan growth is higher a year before the accountability significantly affect loan growth. In the case election, indicating incumbent government's attempt to of Islamic banks in democratic regimes, better socioeco- artificially boost the economy as part of political business nomic conditions and lower internal conflict provides a cycles. Similarly, having loan growth being slower during favourable environment for better loan growth. Corruption election period implies uncertainty of the outcome of the and law and order in the autocratic regime have a negative election with investors using the ‘wait-and-see’ strategy impact on loan growth implying that some amount of cor- (Bjerksund & Ekern, 1990). In general, the findings dem- ruption and lax law and order positively affect loan growth onstrate that governance plays an important role in loan by accelerating and evading bureaucracy, requirements and growth, as all governance variables are significant in the cumbersome rules and regulations. overall sample. For example, corruption is statistically The remainder of the article is organized as follows: negative and significant in all cases except for Islamic the next section reviews some of the recent developments banks in autocratic sub-sample suggesting higher corrup- in Islamic banking along with identifying the identified tion may decelerate loan growth. In addition, in the case principles of Islamic finance. Section 3 aims to present of Islamic banks in autocratic regimes, corruption has a the theoretical framework of this article in the form of positive impact on loan growth suggesting bribery and political economy and institutional economics. It also dis- other forms of corruptions expedite business processes by cusses the articulation of theoretical frames in the form reducing waiting time for approval and bypassing strin- of empirical studies to render a base for the variable gent rules and requirements. Furthermore, regulatory selection. This section also highlights the differences and quality and voice and accountability is found to be nega- recent development in Islamic banking in both demo- tive and significant in all the Islamic banks sub-sample cratic and authoritarian regimes whereby the hypotheses but are positive for the whole sample implying specific are constituted. Section 4 presents the variable definition conventional framework being suitable for the develop- in line with the theoretical framework along with empiri- ment of conventional banking but may retard develop- cal modelling. Empirical analysis and findings are pres- ment in Islamic banking. The length of government ented in Section 5. The final section presents the tenure is consistently statistically negative in the whole conclusions. 4 ASUTAY AND MOHD SIDEK

2 | ISLAMIC FINANCE AND Luxembourg. This positive development is attributed to two BANKING primary reasons. First, the Muslims population in these countries and regions has increased. In Europe, the Muslim Islamic banking emerged to respond to religiously moti- population is around 44 million in 2010, 3.3 million in the vated financial inclusion by providing Shari'ah or Islamic UK or 5.2% of total UK population in 2013 and 6.67 million law compliant financing and banking model for Muslims as or 2.11% of total population in the USA. Second, to tap into an alternative to the interest or riba-based system in the the accumulated wealth emerged due to petrodollars in the conventional banking. Given the paradigm shifts of Muslim GCC region, Shari'ah compliant investment avenues are legalists, methods such as hiyal (ruses) have been used to considered as an important strategy by this non-Muslim get around the theoretical prohibition of riba'resultingin majority countries. Singapore, the UK and Luxembourg, the expansion of financialization through reverse engineer- among others, for example, have gone to a great extent by ing method. As a practice, hence, Islamic banking and providing Shari'ah compliant structures and products, finance complement existing conventional banking offering amending rules to accommodate Islamic finance and pro- products that are within the parameters of the Shari'ah. viding Shari'ah compliant investment avenues in the quest The basic Shari'ah tenets in relation to Islamic to becoming the Islamic financial hub in the respective finance, as part of negative screening, include non- regions. Third, the growth of Islamic finance is fuelled by engagement in interest-related transactions, no financing the consumers' demand for Shari'ah compliant assets by in non-Shari'ah compliant businesses, no investment in Muslims living in non-Muslim majority countries. non-Shari'ah compliant companies, bonds, stocks and Table 1 provides the demographic description of equities, while it promotes profit-loss sharing and risk- the countries from which the sampled Islamic and sharing business models by prohibiting speculative and conventional banks were drawn. Among the countries non-asset-based transactions. with a democratic regime, Malaysia has the highest Despite their contrasting philosophical difference, share of Islamic banking assets, although Muslim pop- Islamic banks are arguably similar to conventional banking ulation is only approximately 63.7% of the total popu- in terms of their operational mechanism (Beck et al., 2013) lation compared to Indonesia, Turkey, Bangladesh and except for Shari'ah compliance, have no significant advan- Pakistan with 87–98% of Muslim population. While tages in efficiency and stability, pro-cyclical, having profit Malaysia ranks lowest Muslim populated country in rate higher than the conventional rates and compete in the the group, but Islamic banking has made important same market as other conventional banking, since the prod- inroads becoming a serious alternative with a signifi- ucts of Islamic banking are basically replications of the exis- cant share in the financial system. Nevertheless, hav- ting conventional products with Shari'ah compliance. ing a large number of Muslims in Turkey, Indonesia Although profit sharing is notably the business models that and Pakistan indicate a potentially significant market fulfil the maqasid al-Shari'ah or the moral objective of for Islamic finance in general despite emerging and Islamic law, fixed income structures, such as murabahah niche Islamic banking sector in these countries. and tawarruq,2 are more successful and constitute large part Among the autocratic regime, the majority of countries of total Islamic banking financing. in our sample have more than 90% Muslim population Except for Iran and Sudan, Islamic banking and indicating a sizeable potential market that has yet to finance is practised as a dual banking within the exis- be tapped. Under such circumstance, it is important to ting (conventional) financial system in the world either understand the role of political institutions in driving as a full-fledged banking sector or as Islamic windows. the development and progress of Islamic finance gener- Shari'ah compliant assets have grown considerably over ally and Islamic banking, specifically. the years despite the total Shari'ah compliant asset accounting only 1.5% of the total global banking assets (The Banker, 2017). Figure 1 shows the total Shari'ah 3 | THEORETICAL FRAMEWORK: compliant assets over 2007–2014 with Islamic banks in POLITICAL ECONOMY, PUBLIC the autocratic regimes having three times bigger than CHOICE, INSTITUTIONAL the share of the total assets. Iran has the most consid- ECONOMICS AND REGULATION erable Islamic banking total assets, as the law and legal SCHOOL framework does not permit the existence of conventional banking system. A number of approaches emerged since the 1960s to In recent years, Islamic finance has permeated into expand the explanatory power of neo-classical economic non-majority Muslim countries such as the USA, the UK, analysis in the face of the observed weakness in its con- Singapore and European countries and especially structs, each of which aimed to render an additional ASUTAY AND MOHD SIDEK 5

TABLE 1 Demographic description: Islamic banking assets

Country Shari'ah compliant assets (USD million) Muslim population % Muslim population Democratic regime Malaysia 204,139 18,929,707 63.7 Indonesia 19,101 217,882,830 87.2 Bangladesh 18,974 140,622,276 89.8 Turkey 13,063 73,433,988 98.0 Pakistan 7,622 175,585,461 96.4 Autocratic regime Iran 516,976 77,214,826 99.7 Saudi Arabia 257,054 26,810,849 93.0 UAE 105,780 7,187,173 76.9 Kuwait 80,425 2,496,112 74.1 Qatar 59,254 1,468,192 67.7 Bahrain 56,593 936,516 70.3 Sudan 7,429 34,433,626 90.7 Egypt 6,532 77,871,503 94.9 Yemen 3,308 24,187,715 99.1 Jordan 1,880 6,016,680 97.2

Note: Compilled with data from The Banker (2013, 2014). strength by bringing omitted paradigms into the theory ‘political economy studies the interdependence between and the modelling of neo-classical economics. Such the economy and the polity of a country or countries’ approaches, among others, including within the larger (Frey, 1978, p. vii; Frohlich & Oppenheimer, 1978). In political economy umbrella, public choice, institutional referring to the endogenizing politics and institutions economics, economic sociology and regulation school. into economic analysis to develop a more effective analy- Their main argument is that economic and financial sis, Caporaso and Levine (1992, p. 31) rationalizes politi- actions, events and policies must be located within the cal economy on the ground that ‘political and economic larger environment by recognizing that politics, political processes and institutions are interlinked and should be institutions, sociological institutions and regulations have studied as a complex and interrelated whole rather than an impact on their emergence, performance and evolve- as separate spheres’. Therefore, public choice theory ment. As it is argued that unlike the simplified version of within the new political economy entirely focused on economic interaction (between households and firms) pro- political and economic interaction to understand the eco- vided by traditional neo-classical analysis, political econ- nomic performance of countries, organizations and indi- omy through its various institutions and in particular vidual economic and political behaviour (Mueller, 1989). through politics and political institutions determine the For example, political business cycle theories explore the functioning of economy and finance. Therefore, in particu- political manipulation of the economy for electoral pur- lar (new) political economy, public choice, institutional pose whereby the change in the business cycles is consid- economics and regulation school attempted to expand the ered as an outcome of rational politicians aiming to be economic analysis through recognizing the role of such re-elected (Alesinaet al., 1997). institutions since the 1980s. The emergence of Islamic eco- Institutional economics (see: Acemoglu, 2008; nomics and finance movement coincides with the same North, 1990) and Regulation School (see: Aglietta, 1976; concern and therefore emerged in the same years in 1960s Boyer, 1990; Boyer & Saillard, 2002) has articulated such and received global recognitions since the 1980s to essen- a political economy concerns and brought the institu- tialize the importance of religion on individual and organi- tional and regulative as well as governance quality to the zational behaviour (Asutay, 2007b, 2013). analysis in an attempt to provide an integrated and com- Political economy can literally be defined as ‘the com- plete analysis. As part of the institutional economics, reli- plexity of interaction between political and economic gion is considered as an important institution behaviour’ (Borooah, 1985, p. 20). In a formal sense, determining individual and economic performance 6 ASUTAY AND MOHD SIDEK

(Barro & McCleary, 2003; Guiso, Sapienza, & Zingales, argued that democracy hinder growth since physical cap- 2003; Iyer, 2010; Kuran, 2004, 2018). ital accumulation is reduced via increased government In line with such developments, as discussed in the consumption spending. Jetter (2014) demonstrate that preceding section, while Islamic economics brought the volatility of government spending has an adverse effect normative values of Islam into economic understanding on growth in democratic regimes but positively affect (Asutay, 2007a, 2007b) since the 1960s, Islamic finance growth in autocratic regimes since in autocratic regimes has produced Shari'ah compliant financial products and the public has limited political involvement, triggering solutions since mid-1970s (see: Asutay, 2015; little or no effect to government spending. In democratic Ayub, 2007). regimes where public voice is one of the determining fac- This article, thus, is located within the larger political tors of government spending, increase in volatility leads economy theoretical framework as expressed through to lower growth. Chen and Liu (2013) show that during public choice, institutional economics and regulation election years in Taiwan, private financial institutions approach. The subject matter of the article is the perfor- have higher loan growth and return on assets. There is mance of Islamic banking which is Islamically deter- also empirical consensus that political instability can mined or Shari'ah compliant institutional form of adversely affect a wide range of macroeconomic variables financing in the sense how the frames (political regimes, such GDP growth (Aisen & Jose Veiga, 2013; Alesina governance quality and political risks) developed from et al., 1996; Jong-a-Pin, 2009), private and public invest- these theories have affected the performance of Islamic ment (Alesina & Perotti, 1996; Darby, Li, & banks. Thus, Islamic banking performance is examined Muscatelli, 2004) and government spending (Devereux & within the interaction of politics, institutions and gover- Wen, 1998) by disrupting long-term economic policies. nance as suggested by the theoretical framework of politi- Importantly, democracy is often argued to be inher- cal economy. The following sections discuss the ently stable due to its predictability, as political stability articulations of these theoretical frames and relevant is vital due to extending policymakers' horizon leading to empirical research in order to develop an effective set of optimal policies although Public Choice school's scepti- variables for this empirical study. cism as to how democracy is used to manipulate the economy leading to welfare losses. In addition, democracy provides transparent rules, which reduce the incen- 3.1 | Political and economic interaction tive of ruling government to implement predatory political decisions on the private resources of the econ- The mainstream economic models view democratic insti- omy. In a democratic system, which encourages open tutions at its best, generate economic growth (Barro, 1989; debate and open public policy decision making, policy North, 1990), while state autonomy is pernicious, the extremism and power taking via illegitimate means can source of inefficiency, impedes growth and development be well avoided. Furthermore, it is argued that democ- (North, 1990). Among others, Baum and Lake (2003) sug- racy reduces policy disequilibrium and is knowledge- gest democracy is an effective tool to ensure maximum induced because the outcome and policy consequences economic freedom, efficient use of resources, and safe- pre and post elections, the aftermath policies and conse- guard the private sphere for the proper functioning of the quences can be well predicted and hence, can be hedged market, which later, trickle down as growth. (Freeman, Hays, & Stix, 2000; Feng, 1997). The existence The empirical impact of the polity on the economy of elections within democratic polity allows more peace- has been relatively well documented; however, studies ful and predictable transfers of governing parties, thus, show mixed results. Przeworski (1966), Huntington and reducing volatility in policies. It can, therefore, be con- Dominguez (1975), Weede (1983), Landau (1986), for cluded that given macro-political stability and ability to example, shows autocratic regimes demonstrates higher make micro-political adjustment existed in democratic economic growth levels. On the contrary, Dick (1974), settings; the economy can grow faster leading to loan and Pourgerami (1988, 1991), Barro (1989) and Feng (1997), deposit growth in banks. amongst others, show that democracy fairs better regard- Within the macro-political settings, institution is con- ing growth. Kohli (1986) and Marsh (1988) show no dif- sidered vital to financial development and growth; and ference between the two regimes. In assessing these two therefore, improvements in the political, legal and eco- institutional frameworks, Butkiewicz and Yan- nomic institutional environment are evidenced to having ikkaya (2006) found that democracy has a negligible positive effects in promoting financial growth (Beck impact on growth, and Helliwell (1992), Tavares and et al., 2003; Huang, 2010; Rajan & Zingales, 2003). For Wacziarg (2001) and Aisen and Jose Veiga (2013) suggest the efficient institutional environment, democratic that democracy has a negative impact on growth. They regime is considered as a priori fostering an open society, ASUTAY AND MOHD SIDEK 7 encouraging fundamental civil liberties and providing a their current strategies to assimilate with the overturned politically stable environment for businesses to prosper. situation. In this case, bankers cannot effectively gather With democracies, protection of property rights and con- enough information or may fail to hedge against the tract enforcement are facilitated which is the key to unexpected changes. Consequently, loan, assets and investment growth (Acemoglu, Johnson, & Robin- deposit growth may be affected in a downward manner. son, 2005; Clague et al., 1996; Olson, 1993). Moreover, it Due to these unexpected political events, bankers need to is evidenced that under democracy, corruption is reformulate their decision and strategies. In addition, suppressed with proper implementation of law and order democracies may not necessarily exhibit smooth political leading to increased efficiency of institutions. Therefore, transition; as elections and constitutional democracy can for the financial market to function efficiently, a robust be superseded by military coup d'etat as in many develop- and inclusive legal and regulatory framework must be ing countries, such as Bangladesh (in 1975, 1981, 2007 present as provided by democracy; as law and order allow and 2011), Pakistan (in 1969, 1977 and 1999) and Turkey for the financial market to operate efficiently with all dis- (in 1960, 1971, 1980 and 1997). putes settled within the legal framework. Democracy, on the other hand, does not always The quality of governance is expected to be superior equate to stability. The in-built feature of democracy in democratic regimes as compared to autocratic regimes. which promotes public participation could create a num- The inclusive political process ensures politicians being ber of pressure groups, which may lead to inefficiency in subjected to regular public scrutiny and checks by the economic decision making. Ruling parties wanting to opposition parties, therefore, minimizing power abuses, preserve their seats in the next election tend to submit to distortionary policies that only benefit a small section of pressures from these groups at the expense of the general the population and control the quality of policymaking. public interests. If lower income groups dominate the Civil engagement in democratic regimes should promote ruling party, political power can be used to raise wages citizen participation in public policies such that special along with redistributive policies which impede invest- attention could be given to marginalized groups to ensure ment and consequently, growth. Countries in a transition financial inclusiveness or access to banking and other from autocracy to democracy may experience internal financial services. Therefore, banking and financial ser- conflict, ethnic problems leading to political instability, vices are available for businesses, individuals and even economic disorder and consequently, slower economic for the majority poor in effective democracies. The exis- and financial growth. Some pressure groups especially tence of a prudent legal framework is vital in ensuring the incumbent firms and government-linked companies the efficacy of contract enforcement, treatment to credi- (GLCs) can directly influence the ruling party to make tors and shareholders, and sound accounting practices. decisions in favour of them despite being in a democratic In democracies, regulation and enforcement is the result regime. of a balance between social and economic constituencies A competing line of argument holds that autocracies (see: Gimma & Shortland, 2008; LaPorta et al., 1997; create political stability in general since long-term poli- Pagano & Volpin, 2001). Mauro (1995) suggests that cies can be carried out without many political distortions. bureaucratic performance such as the level of corruption, The absence of government dissolution and reformation independence and effectiveness of the judicial system of new ruling government, switches between competing and red-tapes affects economic growth, which essential- political parties, which potentially produce unpredictable izes the facilitatory setting of democracies. events and policies changes, can be easily avoided. Bank- In the case of an autocratic regime, many economic ing and financial sector can fully anticipate current and decisions are driven to protect the interest of the elite future political events, since the ruling government is less groups due to restrictions in the participation of type of prone to policy swings, and incorporate this information political system based on patronage. The autocratic in their investment decisions, marketing strategy and regime is thought to impede financial development due future directions of the banks. Stability in the political to lack of suffrage in politics, the heavy influence of elite system is expected to promote overall growth and devel- group over policy and decision making, lack of property opment of the Islamic banking system. rights laws and lack of implementation of law and order. The second distinct feature of authoritarianism is its Arguably, elections and polls in democratic regimes capacity to insulate and control the effects of pressure may give rise to uncertainty. In the presence of two con- from influential firms and unions. Individual firms' trasting competing parties, the outcome of elections may objectives are often suboptimal, short-term and specific result in the dissolution of the current ruling party and to their firms, unions or production, which consequently replaced by its opposition. If their policy direction lead to underinvestment, un-optimal use of resources and changes starkly, bankers may need to revise or revamp less efficient production. In an autocratic regime, the 8 ASUTAY AND MOHD SIDEK state is free from distributive pressures emanating from cronies of the top officials and groups of individuals the public, which facilitates implementation of growth- financially assist them during previous elections. In this enhancing measures. State autonomy promotes economic case, an electoral democracy assimilates features of growth via institutions, education, increasing returns to autocracy when a group of elites control the parties or scale and allocative efficiency (see: Haggard, 1990; the electoral agenda. Elites groups, which control a sig- Przeworski & Limongi, 1993; Wade, 1990), since political nificant proportion of the economy, may exist in a demo- autonomy permits the state to extract resources effi- cratic system, who normally have strong back-up from ciently, provide public goods, and incurs necessary short- the government. Government support towards these elite term costs to facilitate economic adjustments and other comes in the form of costly private sector projects backed growth-related policies. by a government, GLCs, ‘letters of support’ or ‘unwritten In an autocratic setting, the role of public opinion in guarantees’ which has legal implications. These elite influencing banking policies is limited by definition. The groups will later finance electoral campaigns and other absence of a channel through which public can express activities to secure tenure-ship of the ruling party. The their desired level of banking services is expected to existence of these elite groups is more evident during lower growth rates of the banking system in autocratic bailout exercises in the aftermath of a crisis. Therefore, regimes. Many autocratic regimes experience violent and despite having substantially different philosophic founda- erratic political transitions. Examples from the develop- tions, democratic and autocratic rulers may express ing world include Egypt in 2011 and 2013, Iran in 1921, themselves in the same manner in policy making and dis- 1953 and 1980, Sudan in 1958, 1969, 1985 and 1989, and tribution of resources. Yemen in 1962 and 1974. Political turmoil arising from It is sometimes hard to relate political regimes with military coups and hostile takeover exert harmful effects economic or financial development. Autocratic regimes on growth. have different economic objectives and policies, which Acemoglu (2008) further argues that oligarchies fail can later impact their economies differently (see: De Long to take advantage of new growth opportunities due to the & Shleifer, 1993; Jones & Olken, 2008; Larssons & erection of barriers to entry especially entry of new entre- Parente, 2013). As such, it is sometimes argued that the preneurs and new technology, which later, undermines objectives of the government are the main determining growth potential. In his theoretical exposition, factors towards economic and financial growth, rather Acemoglu (2008) shows that oligarchies protect the prop- than which political regime they choose to reign. Reasons erty rights of the producers and prevent distortionary for this include democracy having a comprehensive con- taxes, which in turn enable politically powerful elite cept (Persson and Tabellini, 2006) and the distinction groups to monopolise. In contrast, democracy promotes a between democratization and economic and financial more egalitarian distribution of resources by suppressing liberalization. the power of incumbent elites. On the other hand, an oli- Financial liberalization is vital in the case of banking garchic government may have pro-business policies, if such liberalization includes granting of new banking which spur the economy to unprecedented growth; South licenses, encouraging foreign banks to enter the domestic Korea and Japan in the 1960s are the ‘living proof’ of this market, more investor-friendly rules and regulation. This intuition. might require some control over the economy; and there- As more countries in the world are aiming to consoli- fore, economic and financial liberalization in developing date their democracy, we also witness the elements of countries mainly took place under autocratic regimes autocracy within a democratic setting especially when (see the cases with Argentine, Turkey, Indonesia, among the ruling party has more than two-thirds majority, such others). On the other hand, Aidt et al. (2008) and, to as Malaysia (1955–2004) and Singapore (1959–2011). some extent, Flachaire et al. (2014) suggest that institu- Democratic rulers have the incentive to increase the tions that shape the policy framework and implantation sphere of governmental activities to maximize their lever- of those policies, corroborating the relevance of the auto- age over the economy and to ensure the durability of cratic versus democratic regime debate. their tenures. The elements of property rights as On theoretical grounds in institutional economics, championed under democracy is further enhanced with corruption is dreaded, since it distorts the market econ- not only property rights ensured to politically linked con- omy and the efficient running of the financial system glomerates, but also secured internal market (see: resulting in reduced efficiency of government and busi- Acemoglu, 2008). Autocratic elements within a demo- nesses since corruption allows certain elite members to cratic structure can be seen via the existence of elite gain businesses through patronage rather than ability. groups. These elite groups are primarily related to the Bribes or special payments make the cost of running ruling government via political and non-political ties, business higher and consequently; the investment may ASUTAY AND MOHD SIDEK 9 be held back. In contrast, some businesses are willing to different regions and that such behaviour is indifferent pay bribes to expedite the business procedure if the cost regardless of being in a democratic or autocratic regime. of the bribe can be offset by the profits gained. In this Cole's (2009) study on India shows that bank credit is 5– case, corruption has a positive effect on the economy, as 10% higher in election years, especially in districts, which evidenced by the empirical support offered by are heavily contested. On the contrary, in examining a Mauro (1995). sample of commercial, foreign and state banks from Tur- key for the period of 1967 to 2007, Baum et al. (2010) found that elections have significant effects on bank 3.2 | Political-institutional environment behaviour in the form of loan growth, deposit growth and banking sector nexus and bond growth but they could not find evidence for elections leading to higher lending rate whether the As regards to the impact of political institutions and their banks are state-owned, domestic or foreign-owned nature on financial and banking sector, this can be banks. Similarly, Onder and Ozyildirm (2013) in their viewed from several perspectives. It is argued that gov- study covering 67 (1992) to 81 (2000) provinces ernments can significantly impact the financial activities between 1990 to 2010, conclude that state-owned not only via policies but through indirect political influ- banks give credit for political reasons during election ences on major financial institutions especially govern- years as results from their study show that the share ment-linked financial institutions or politically of state-owned banks in Turkey during crisis and elec- connected individuals hold important decision-making tion years are significantly higher. Chen and Liu (2013) positions in the financial institutions. Government- found that private banks earn higher ROA and loan owned banks, for example, may channel funds to politi- growth during election years compared to foreign- cally motivated projects while disregarding public inter- owned banks in Taiwan from 1994 to 2009. At the est (Baum et al., 2010; Chen & Liu, 2013). same time, government-owned institutions lending The political channel is expected to be stronger in were no longer affected by political pressures denoting countries with democratically elected members. During the success of financial reforms undertaken by the election years, for example, bank lending is expected to Taiwanese government in 2002. increase substantially through supply shock or demand Bongini et al. (2002) show how government interven- shock channels. If bank lending which leads to increase tion motivated by political connections help to rescue on loan growth is accompanied by a decrease in the cost deteriorating banks. Unqualified borrowers with political of lending, then we hypothesize that lending is driven by connections get to borrow from state-owned banks or supply shocks or what is termed as ‘political lending’ banks whose shareholders have political connections (Micco et al., 2007). On the other hand, demand shock is (Fraser et al., 2006; Khwaja & Mian, 2005). A study in where bank lending is associated with an increase in the Ukraine by Baum et al. (2008) suggest that interest rate price of the loan. From a political economy perspective, margin is much lower in politically affiliated banks. On a banks may increase their lending to state-owned enter- similar note, state-owned banks tend to charge lower prise or to firms with politically connected directors interest rates in areas where these banks have a strong (Khwaja & Mian, 2005; Sapienza, 2004) in an attempt to affiliation with the political party, which won the election garner more support from the corporate sector for the (Sapienza, 2004). In similar veins, Jackowicz et al. (2013) elections for favourable returns. On the contrary, auto- argue that the net interest margin during election years is cratic leaders may impose their will on banks especially lower due to politically motivated policies. In their study, state banks to achieve specific political or economic goals which covers banks in Central Europe, lower net interest (see among others, Gabgub, 2007 on the impact of Libyan margin is the outcome of a lower interest rate on loans. government on bank lending). In examining the impact of political uncertainty, The work by Dinc (2005), Micco et al. (2007), Flachaire et al. (2014) found that greater political uncer- Sapienza (2004) and Khwaja and Mian (2005) provide tainty is associated with higher cost of loans: the esti- empirical evidence where state banks increase loans dur- mated increase in interest spread is 11.90 basis point for ing election years, whose findings are robust across dif- one standard deviation increase in idiosyncratic political ferent time period and sample sizes. Sapienza (2004) and exposure. Khwaja and Mian (2005) studies Italian and Pakistani A more recent study by Hernandez and banks, respectively, while Dinc (2005) and Micco Vadlamannati (2016) shows Saudi Arabia's political influ- et al. (2007) extended the study to a more extensive set of ence on the Islamic Development Bank favouring the countries. Micco et al. (2007) deduce that increased lend- Sunni-school following Muslim states vis-à-vis the Shia ing behaviour during election years are not affected by majority states when giving out loans. 10 ASUTAY AND MOHD SIDEK

Eichler and Sobanski (2016) show that electoral financial system and the political regimes cycle, the ideology of the government party and the (Rethel, 2010, 2011) and therefore rather than Islamic power of the government significantly affect the stabil- governance of extended stakeholding, shareholder value ity of banks in the Eurozone. Specifically, government maximization determines the operational nature of cor- tenure, government majority in parliament, govern- porate governance in Islamic banking. ment fractionalization and left to right party preference While AAOIFI (Accounting and Auditing Organisa- positively and significantly affect bank stability. Their tion for Islamic Financial Institution) and IFSB study also suggests that smaller banks are more sub- (Islamic Financial Services Board), as the two main jected to national politics compared to larger banks standard setters for the industry, have issued a number which are highly capitalized which makes them unaf- of standards in relation to the operation and the gover- fected by risky political environment. In the case of nance of Islamic banks (Asutay, 2015), such standards the UK, Kleymenova et al. (2016) show that after still remain within the frames of the existing financial nationalization, non-British banks increased their exter- system rather than disrupting them. Therefore, as men- nal funding and locate their lending away from the tioned above, Islamic banks represent hybrid institu- UK in the sample of 334 banks in the UK from tions with their hybrid products (Asutay, 2012) to 1999Q1 to 2011Q4. Ashraf (2017) shows that political ensure efficiency and competition in the dual banking and legal institutions affect banks risk-taking behav- system by using Islamic norms in the form of Shari'ah iour. In a sample of 98 countries ranging from 1998 to compliance. 2007, he shows how healthy political institutions boost Based on the existing literature discussed so far in the credit market competition due to increased risk-taking preceding sections, political and institutional factors as behaviour but would generate moral hazard problems the articulation political regimes seem to have some since there is some assurance for the government bail- impact on the banking system and the performance of out in the case of default. banks. However, due to being hybrid institutions within the capitalist economy, Islamic banks are expected to be affected by the same political and institutional factors. 3.3 | Islamic banking and political However, research on how the political system and their economy interaction: Hypothesis respective institutional environment affect Islamic bank- development ing is almost absent except for Bitar et al. (2017). In their study, Bitar et al. (2017) aim at examining the impact of While the initial discourse of Islamic economics democratic and non-democratic political environment suggested an economic system generated through the and also the impact of legal systems (Shari'ah based, ontological base of Islam and its value system and nor- hybrid and Western) on the financial soundness of mative principles (Asutay, 2007a, 2007b, 2012), due to the Islamic banks. They conclude that compared to the finan- global political economy this was not possible (Elashker cial soundness of conventional banks, Islamic banks & Wilson, 2006). Hence, Islamic banks have been operat- underperformed in the democratic political environment, ing within the existing political and economic system as while performed better than conventional banks in hybrid institutions providing hybrid products. In other hybrid and Shari'ah based legal system. words, except for Iran and Sudan, Islamic banking has This article intends to fill the observed gap in the lit- been operating within the dual banking system under the erature by presenting an extended analysis using differ- impact of the existing political regimes and institutions ent proxies to capture the effects of political regime and and mostly under the existing banking law and regula- political risk as well as governance quality on Islamic tions. While Iran and Sudan provide Islamic political banks and conventional banks in the sampled countries environment for the operation of Islamic banks, in recent in a comparative manner. To operationalize the research, years several countries, led by Malaysia, issued special in light of the discussion provided so far, the following law and regulation for the efficient operation of Islamic hypotheses are developed: banks. However, Islamic banking laws and regulations still essentialize the neo-liberal economy and financial H1 Democratic regimes (in various forms) have a posi- environment (Rudnyckyj, 2013, 2014) rather than devel- tive impact on financing/loan extended by Islamic oping an alternative system. Therefore, within the hege- banks. mony of the existing financial system, Islamic banks must comply with Basel rules in order to be competitive. H2 In democratic countries, elections have a positive In a similar way, the corporate governance of Islamic impact on financing/loan extended by Islamic banking institutions is determined by the existing banks. ASUTAY AND MOHD SIDEK 11

H3 Quality of political governance has a positive impact chief executive. The concept variables provide an alterna- on the financing/loan extended by Islamic banks. tive approach towards understanding polities of the authorities since there are various paths to democracy H4 Political risks have a detrimental impact on the and autocracy. The election years were obtained from financing/loan extended by Islamic banks. Database of Political Institutions (DPI). An alternative proxy for the political regime is the The political and institutional constituents of these database provided by Freedom House, which is divided hypotheses are discussed in detail in the following sec- into two major categories political rights (pr) and civil lib- tions to demonstrate the choice of representative and erties (cl). The combination of these variables determines appropriate variables as listed in Table A2. whether a country is free, partly free or not free. Political rights (pr) have three major components: electoral process (a), political pluralism and participation (b) and 4 | VARIABLE DEFINITION, functioning of the government (c). Civil liberties (cl) cap- SAMPLE AND ECONOMETRIC ture the freedom of expression and believe (d), associa- SPECIFICATIONS tional and organizational rights (e), the rule of law (f) and personal autonomy and individual rights (g). The use 4.1 | Variable selection and definition of these three different databases allows us to capture different aspects of political regimes. In overall, the Polity Based on the theoretical framework and its articulation IV database captures details of the inside mechanism presented in Section 3 as well as the discussed empirical within the institutions, while the Freedom House data- literature, three sets of institutional variables were con- base assesses the overall running of the political sidered to examine the impact on loan or financing institutions. growth in Islamic banking in the dual banking system The second part of the institutional variable is gover- with the additional of Islamic banks in the Islamised nance. The World Governance Index (WGI) issued by the economies of Iran and Sudan. These are political regime, World Bank provides seven measures to capture the effi- governance quality and political risk. Political regime cacy of governance worldwide: control of corruption related variables refer to Hypothesis H1 and Hypothe- (cor), government effectiveness (g_eff), political stability sis H2, governance quality related variables aim to test and absence of violence and terrorism (p_stab), regula- Hypothesis H2 and Hypothesis H3 is tested by political tory quality (reg), rule of law (rol), voice and accountabil- risk factor. By considering these variables, political econ- ity (voa) and tenure (ten). Control of corruption (cor) omy theoretical framework's aim of endogenizing the examines to what extent public power is exercised for the impact of politics and institutions is endogenized in the private gain of elites and private interests. Government study. The following section presents the selection and effectiveness (g_eff) describes the quality of public and definition of these variables to provide a further ratio- civil services, independence from political pressures, pol- nale, while a summary of variable definitions and sources icy formulation and implementation and government's of data is provided in Table A2. commitment towards realizing the policies. Political sta- First, political regimes and their features are captured bility and absence of violence (p_stab) captures the possi- using two databases: Policy IV and Freedom House. Pol- bility of overthrowing governments in an ity IV variables consist of the degree of democracy (demo) unconstitutional manner. Voice and accountability (va) and autocracy (auto). In the case for a democratic regime, shows freedom of expression, freedom of association, free six polity component variables and three concept vari- media and ability to participate in the selection of gov- ables are included. The extent of institutionalization of ernment freely. The rule of law (rol) captures the extent executive transfer (xrreg), the competitiveness of execu- of contract enforcement, property rights, the likelihood of tive selection (xrcomp), the openness of executive recruit- crime and violence and how much confidence agents ment (xopen), executive constraint (xconst), degree of have in abiding the rules of the society. Finally, regula- organization and institutionalization of selection (parreg) tory framework (reg) summarizes the ability of the gov- and degree to which political participation is free from ernment to formulate and implement sound policies and government control (parcomp) represent the polity com- regulations that will enhance private sector development. ponent variables; executive recruitment (exrec), executive In short, the process by which government are selected, constraint (exconst) and political competition (polcomp) monitored and replaced are represented by voice and capture the concept variables. These variables include accountability (va) and political stability (p_stab). The factors ranging from the degree of competitiveness of efficacy of government policy formulation and imple- political participation to the degree of constraints on mentation is captured by government effectiveness 12 ASUTAY AND MOHD SIDEK

(g_eff), and regulatory quality (reg) and the confidence of Goemans (2010) generally classify the countries in their the people towards the state is indicated by the rule of study into democracy and authoritarian regimes. Author- law (rol) and control of corruption (cor). itarian regimes are then categorized into monarchies, Thethirdpartoftheinstitutionalvariableisthe military regime and civil dictatorship. Tongur et al. (2015), political risk. Data is obtained from the International for example, classify democracy into a social democracy, Country Risk Guide (ICRG), from the PRS group. We conservative democracy and one-party democracy while use six sub-components of the ICRG political risk mea- autocracy is divided into dictatorship and military dicta- sures: government stability (g_stab), socioeconomic torship. However, Collins (2006) argue that political conditions (socio), internal conflict (int_con), demo- power distribution varies amongst countries which cratic accountability (dem_acc), corruption (cor2)and requires a more restrictive definition. Accordingly, law and order (lao). Only 6 out of 12 measures were Malaysia is classified as a competitive authoritarian selected owing to their applicability to the countries in regime, while UAE is considered as a tribe-based authori- our sample. Among these, government stability tarian regime, while most of the studies consider Malay- (g_stab) examines the ability of the ruling government sia democratic through a functional definition. In to stay in office and fulfilling their election manifesto. summary, there are different arguments as to how they The risks are calculated based on three sub-compo- classify the countries into different regimes. nents namely government unity, legislative strength In this article, selected Muslim countries are classified and popular support. Socioeconomic conditions (socio) into autocratic and democratic regimes only which is capture the risks of social dissatisfaction arising from mainly dictated by the availability of data. Accordingly, a socioeconomic pressures with unemployment, poverty country is categorized as democratic if the polity2 score is and consumer confidence as sub-components. Internal higher than 5 and holds regular elections. Otherwise, the conflict (int_con) assesses the risk of civil war, civil dis- countries are categorized as autocratic. This division order and political violence that can potentially harm allows us to examine the overall and specific impact on the governance. This also includes whether the ruling Islamic banks in both regimes. Further disaggregation of government engage in arbitrary violence against its countries under autocratic and democratic regimes people, armed or civil opposition to the government through the articulation of particularities leads to fewer and on-going civil war. Corruption (cor2)examines observations, which may affect the efficiency of the corruption in the political system. Three of institu- estimation. tional political variables, namely, corruption (cor and cor2), the rule of law (lao, rol and rule of law, f)and political violence (int_con and p_stab)fromthethree 4.3 | Sample different databases seem to overlap. Since the measure- ments and definitions are entirely different, this can be The analysis in this article includes a sample of 16 Mus- considered as a robustness test. lim majority countries from 2000 to 2013. The choice of countries and time range is substantially limited by the availability and the consistency data. Asa can be seen in 4.2 | Classification of countries Table A1 (see Appendix), a total of 138 banks were according to the regime type crossed examined. This sample consists of 69 conventional banks and 69 Islamic banks with 54 banks from The main issue in the empirical construct is how to disag- autocratic regimes and 15 banks from democratic gregate the countries into regime types to facilitate a regimes. The sample is divided into four categories: con- meaningful empirical analysis. Such aggregation is aimed ventional and Islamic banks, Islamic banks only, Islamic to illustrate the different requirements to maintain their banks in the democratic regime and Islamic banks in power in those countries, which is later translated into autocratic regimes. how the country is being managed and policies are The sample only includes banks with 5 years or more undertaken. Under a democratic regime, we would observations to capture the impact of politics since expect the elected party(ies) to adopt policies that would shorter time span may not be able to capture such an benefit the majority of the citizens in the country to institutional impact. In the sample, democratic countries ensure their political survival. Leaders in democratic include Turkey, Malaysia, Indonesia, Bangladesh and regimes are presumed to avoid costly short-term policies Pakistan. GCC countries (Bahrain, Kuwait, Oman, Qatar, to safeguard their tenure. Cheibub et al. (2010), Steinberg Saudi Arabia, United Arab Emirates), Yemen, Egypt, and Malhotra (2014), Hadenius and Teorell (2007), Iran, Sudan and Jordan represent the autocratic regime- Magoloni (2006), Lai and Slater (2005), Debs and based countries in the sample. ASUTAY AND MOHD SIDEK 13

4.4 | Estimation method the institutional and economic effects on loan growth canbeeitheroverorunder-estimated.Sincethenum- Our empirical model is based on Baum et al. (2010) ber of external instruments to address all potential which takes the following form: endogenous variables are limited, GMM relies on the lags of the regressors as instruments circumventing α γ φ τ ε Loan growthit = 0 + Bit + Eit + Mit + vi + it endogeneity and reverse causality bias. where i and t denote bank and time, respectively, and γ, τ and φ are vectors of coefficients on the bank-specific fac- 4.5 | Preliminary statistics tors, crisis dummies and institutional factors. vi captures the banks fixed effects, and εit denotes the error term. Table 2 presents descriptive statistics for the primary vari- The elements of B control for bank-specific factors which ables, while Table A3 (see Appendix) provides autocorre- may affect loan growth. Total assets (ta) is used to control lation results. As can be seen, none of the variables is for the size of the bank and the effect of economies of highly correlated, which allow bundling of sub-categories scale. Deposits over total loan (dtl), a proxy for liquidity, variables according to political regimes, governance and is included to measure the density of financial strength political risks in regressions. The countries selected are that could be used to give out loans. In addition, a set of vastly different in terms of size, political ideology, cul- institutional data (E) are used to represent the political ture, and income, which may potentially lead to non-sta- regime, governance and political risks. Finally, a crisis tionary heteroscedasticity. To deal with this issue, we rely dummy (M) is introduced to capture the macroeconomic on panel unit root tests proposed by Herwartz and effects. Siedenburg (2008), tHS, and Demetrescu and Based on initial inspection of the data, there are varia- Hanck (2012), tDH, which caters for cross-sectional tions across the institutional variables within each coun- dependence and non-stationary heteroscedasticity. try. Even within the autocratic and democratic regimes, Table A4 (see Appendix) presents the HS and DS test sta- the degree of autocracy and democracy varies signifi- tistics, which shows that the data is stationary at level. cantly across the sample of countries, prompting the use Additionally, the usual panel unit root tests, LLC, IPS, of fixed-effect panel regression models. The autocratic ADF-Fisher and PP-Fisher are also applied to the regime in the sample comes in the form of autocracy, oli- variables.3 garchy and monarchy. Following institutional changes Unit root test results show that all variables except due to the political upheaval following Arab Spring in demo, auto and g_stab are stationary or I(0) in at least 2011 and the possibility that the effect will proliferate to one of the tests, corroborating the use of the basic panel neighbouring Arab countries, we resort to the use of data models rather than the PMG-type models which fixed-effect method. Fixed-effect approach is used to con- deal with stationarity. However, to introduce dynamism trol for country-specific time-invariant traits that are not into the equation, we introduce lagged values for the accounted for by the control variables in our empirical bank-specific effects. For robustness check, we employ model. feasible GLS and GMM. To introduce dynamism in the static model, we use lagged variables in two control variables: total asset (ta) and deposit over total loan (dtl). For robustness 5 | POLITICAL REGIMES, check, we employed GLS (generalized least squares) GOVERNANCE QUALITY AND and GMM econometric models. GLS with AR(1) cor- POLITICAL RISK, AND LOAN rection is employed to accommodate the effect of auto- PERFORMANCE OF ISLAMIC correlation and heteroscedasticity (Phillips BANKS: FINDINGS et al., 2013). In addition, GLS is more efficient compared to feasible GLS in small to medium size sample. The empirical analysis and results in this section are However, estimation using fixed-effect approach and divided into two parts. The first part reports the impact of GLS may be biased due to the introduction of lagged political regime, governance and political risks on financ- dependent variables among the regressors, which may ing or loan growth. In the second part of the results, we lead to reverse causality. To partly assuage this prob- introduce interaction variables between loan growth and lem, we supplement the empirical results with GMM all the institutional variables. These interaction variables estimation (Blundell & Bond, 1998), which aims at are expected to capture the short-run impact of institu- tackling the problem of endogeneity and reverse cau- tional behaviour on loan growth and their marginal sality. Depending on the source of endogeneity bias, effects. 14 ASUTAY AND MOHD SIDEK

TABLE 2 Descriptive statistics of main variables: Overall The hypothesis that political institutions matter for sample loan growth has explicit empirical support which pro-

Variable Mean SD Min Max vides evidence for H1. Polity2 variable is both positive and significant. Tables 3 and 4 show that political regime loan growth 0.0071 0.0170 −0.0815 0.1758 matters for loan growth. As the degree of democracy size 6.8820 1.1531 3.3152 8.9621 (demo) improves, loan growth is positive and significant. liquidity 0.4732 0.2302 0.0119 0.9903 Similarly, as the degree of autocracy increase, loan polity2 −5.7556 3.1870 −10 3 growth is negative and significant implying lower growth demo 0.4075 0.9221 0 4 rates in the autocratic regime compared to a democratic auto 6.1630 2.5028 0 10 regime. This is indicated by the positive sign of polity2 and demo, while auto is negative in Table 3. Thus, H1 is pr 5.9259 0.8794 4 7 supported by some of the findings; however, when it cl 6.1323 0.8079 4 7 comes to the more specific aspects of the political regime cor −0.3125 0.8390 −1.5066 1.7228 variables, the results demonstrate distinctions. As can be g_eff −0.3309 0.7851 −1.4577 1.1379 seen, three of the polity component variables are negative p_stab −0.9062 1.1502 −2.6600 1.2125 and significant. More standardized procedures for execu- reg −0.4154 0.9231 −1.7305 1.1161 tive power transfer (xrreg), more competitive selection of executive (xrcomp) and more regulated participation rol −0.3762 0.8666 −1.5984 1.0324 (parreg) leads to slower loan growth. When the sample is va −1.2220 0.4333 −1.8568 −0.1897 reduced to Islamic banks, only executive power transfer ten 12.6369 7.8458 1 33 (xrreg) and regulated participation (parreg) is negative g_stab 0.7680 0.1352 0.4167 0.9583 and significant. Taking into account Islamic banks in socio 0.4574 0.2193 0.125 0.9167 democratic countries, executive transfer (xrreg) and selec- cor2 0.4674 0.2757 0 1.1667 tion of executive (xrcomp) is negative and significant. lao 0.4446 0.2830 0.1667 0.8333 Executive constraint (xconst) is marginally negative and significant in democratic regime cases, which means that dem_acc 0.5887 0.2023 0 1 accountability group (for example, legislator and judi- int_con 0.6797 0.1515 0.0000 1 ciary) has the power to limit executive (for example, pres- ident, king, cabinet) decisions. The negative relationship between xconst and loan growth implies that with a lim- 5.1 | The relationship between loan ited constraint on the executive, loan growth can grow growth, political regime, governance and more as oppose to if the accountability group overrides political risks the executive decisions. As for the concept variables, executive recruitment The econometric analyses are divided into four sub- (exrec) is positive and significant across all samples. This samples: overall sample which covers all banks in con- suggests that more institutionalized, open and competi- ventional and Islamic banking, Islamic banks only, tive mechanisms of political leader selection promote Islamic banks in the democratic regime and Islamic positive loan growth. The competitive selection that banksinanautocraticregime.Theresultofthecontrol results in leaders that are competent and pro-business variables generally conforms to the theoretical predic- will device growth-enhancing policies, which eventually tion. Log of total assets (size) and deposit as a ratio of lead to more businesses. As the results evidence, execu- the total loan (liquidity) are used to control for the size tive constraint (exconst) is negative and significant in all and liquidity of the banks, respectively. Size of the samples. Results imply that less constraint on executive bank (ta) is positive and significant in all regressions authority leads to higher loan growth as decisions can be for the overall sample, implying bigger banks are capa- implemented with minimum interference. Therefore, ble of offering more loans. For the sub-categories, only pro-business policies can be undertaken effectively with a few regressions have a positive and significant effect. minimal political intervention. Nevertheless, size is still positive for all other regres- The results of the extended model which uses data sions albeit being insignificant. While results, hence, from an alternative source are reported in Table 4. Politi- imply that higher liquidity (liquidity)promoteshigher cal rights (pr) and civil liberties (cl) are both negative and loan growth, liquidity is both positive and significant significant indicating that loan growth is positive as in overall sample but becomes insignificant in the sub- countries move more towards democracy. All the sub-cat- samples. egory variables for political rights are significant. As the ASUTAY AND MOHD SIDEK 15

TABLE 3 Impact of political regime on loan growth (Dependent variable: Loan growth)

Overall sample (Islamic and Conventional banks) Islamic banks

Variables 1 2 3 4 5 6 7 8 9 10 size 0.5458*** 0.8375*** 0.5461*** 0.9078*** 0.7180*** 0.0010 0.0009 0.0011 0.0022** 0.0025** (0.0944) (0.0928) (0.1125) (0.0929) (0.0976) (0.0009) (0.0009) (0.0009) (0.0010) (0.0010) liquidity 0.0494** 0.1653*** 0.0267 0.1785*** 0.5612*** 0.0101*** 0.0097*** 0.0107*** 0.0135*** 0.0153*** (0.0225) (0.0224) (0.0274) (0.0238) (0.0519) (0.0036) (0.0037) (0.0036) (0.0040) (0.0040) c 4.6602*** 7.3672*** 3.8298*** 11.3982*** 3.3380*** 0.0116 0.0082 0.0149* 0.0299** 0.0214** (0.6257) (0.6268) (0.7487) (1.1940) (0.8975) (0.0085) (0.0082) (0.0088) (0.0137) (0.0097) crisis −3.1052*** −1.1196 −2.8649*** −2.2039*** −0.0207** −0.0199* −0.0206** −0.0180* −0.0178* (0.8776) (1.0230) (0.8467) (0.8295) (0.0104) (0.0104) (0.0103) (0.0105) (0.0106) polity2 0.3225*** 0.0003* (0.0158) (0.0001) demo 0.4013*** 0.6248*** 0.3740*** 0.0003 0.0013 0.0011 (0.0185) (0.0779) (0.1410) (0.0003) (0.0020) (0.0013) autoc −0.1824*** −0.0006** (0.0311) (0.0003) xrreg −0.6242*** −0.0029** (0.1831) (0.0013) xrcomp −0.6231*** 0.0020 (0.2178) (0.0027) xopen 0.1663 0.0005 (0.1083) (0.0009) xconst −0.2247 −0.0009 (0.1643) (0.0011) parreg −0.6357*** −0.0019* (0.1553) (0.0011) parcomp 0.1386 0.0003 (0.1582) (0.0014) exrec 0.5887*** 0.0024*** (0.0843) (0.0008) exconst −0.6166*** −0.0033*** (0.1741) (0.0011) polcomp −0.0064 0.0003 (0.0630) (0.0004) obs 1,012 1,012 1,012 1,012 782 612 612 612 612 590 Within R2 0.5685 0.5837 0.4063 0.6190 0.6573 0.0328 0.0404 0.0360 0.0403 0.05000 Between R2 0.2902 0.2674 0.2198 0.4492 0.5866 0.0306 0.0363 0.0280 0.0361 0.0384 Overall R2 0.5113 0.4711 0.4016 0.5512 0.6119 0.0673 0.0749 0.0693 0.0752 0.0809 F-test 3.21 3.85 6.34 3.31 3.58 4.68 4.58 4.72 4.60 4.55 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Islamic banks in democratic regime Islamic banks in autocratic regime

Variables 11 12 13 14 size 0.0214*** −0.0224*** −0.0013 −0.0023** (0.0067) (0.0081) (0.0010) (0.0010) liquidity 0.0618*** 0.0760*** 0.0120*** 0.0142*** (0.0231) (0.0281) (0.0039) (0.0038) c 0.3110*** 0.1502 0.0140 0.0071 (0.0846) (0.0907) (0.0120) (0.0113) (Continues) 16 ASUTAY AND MOHD SIDEK

TABLE 3 (Continued)

Islamic banks in democratic regime Islamic banks in autocratic regime

Variables 11 12 13 14 crisis −0.0102 −0.0006 −0.0002 −0.0002 (0.0186) (0.0202) (0.0023) (0.0022) xrreg −0.0593** −0.0081*** (0.0228) (0.0031) xrcomp −0.0083** 0.0154** (0.0042) (0.0062) xopen −0.0137 −0.0020* (0.0098) (0.0011) xconst −0.0052* −0.0001 (0.0027) (0.0014) parreg −0.0084 −0.0007 (0.0063) (0.0018) parcomp 0.0078 0.0002 (0.0069) (0.0014) exrec −0.0108** 0.0032** (0.0051) (0.0013) exconst −0.0153*** −0.0015 (0.0047) (0.0012) polcomp 0.0001 0.0002 (0.0030) (0.0006) obs 138 112 478 478 Within R2 .1899 .1750 .0461 .0489 Between R2 .4325 .2886 .7274 .3954 Overall R2 .0850 .1112 .0586 .0707 F-test 3.66 3.61 3.40 3.10 (0.000) (0.000) (0.000) (0.000)

Notes: ***, ** and * represent 1%, 5% and 1% significance level. Robust standard errors are in parentheses for the coefficients. p-Values are in parentheses for F-statistics. results show, free election process (a) and political plural- effect in almost all regressions. One possible explanation ism and participation (b) promotes positive loan growth is that Islamic banks are too small to be impacted by the in the overall sample. Functioning of the government, crisis and the use of non-varying or fixed rate cushioned however, does not necessarily lead to higher loan growth. the effect of crises along with the fact that the business The sub-categories are not significant in the sub-samples cycles in the sampled countries were not profoundly and except for functioning of the government in the case of adversely affected by the global financial crisis. Similar Islamic banks in democratic regimes. As evidenced in the findings by Ghosh (2015) suggest that the Arab Spring results, three civil liberties (cl) sub-components are signif- did not affect loan growth nor political transition. icant: freedom of expression, free media (d), and more The results demonstrate that the election dummy var- associational and organizational rights (e) promote loan iables are negative and significant in the overall sample, growth in the overall sample. The rule of law, on the which implies that elections create some kind of uncer- other hand, is negative and significant implying lax legal tainty for bankers and potential customers, which affects enforcement is beneficial towards loan growth. Similarly, loan growth. As for Islamic banks in democratic regime the sub-categories are insignificant in the sub-samples. sub-sample, election variable is negative, but pre-election The impact of the crisis dummy variable (crisis) is dummy variable is positive, which indicates the classic both negative and significant in the overall sample. How- political business cycle argument where incumbent gov- ever, when the sample is split into other sub-sample, the ernment portrays right economic conditions prior to the crisis is no longer significant but retained the negative election through manipulating the economy. Thus, ASUTAY AND MOHD SIDEK 17

TABLE 4 Impact of political regime on loan growth (Dependent variable: Loan growth)

Overall sample (Islamic and Conventional banks) Islamic banks

1 2 3 4 5 678910 size 0.6061*** 0.8359*** 0.7215*** 0.8704*** 0.8413*** 0.0012 0.0002 0.0009 0.0007 0.0008 (0.0966) (0.0945) (0.0930) (0.0964) (0.0922) (0.0009) (0.0009) (0.0009) (0.0010) (0.0009) liquidity 0.1021*** 0.0648*** 0.0972*** 0.0777*** 0.1614*** 0.0120*** 0.0126*** 0.0110*** 0.0160*** 0.0099*** (0.0231) (0.0244) (0.0222) (0.0245) (0.0224) (0.0037) (0.0042) (0.0037) (0.0042) (0.0037) c 10.4504*** 5.9104*** 11.2417*** 4.5240*** 7.7083*** 0.0040 −0.0058 0.0058 0.0060 0.0086 (0.7122) (0.5887) (0.6903) (0.5860) (0.6286) (0.0083) (0.0085) (0.0084) (0.0092) (0.0083) crisis −1.0759 −2.2641*** −3.6657*** −0.0183* (0.8948) (0.8632) (0.8660) (0.0103) demo 0.4258*** 0.0003 (0.0195) (0.0003) pol_rights −1.2589*** 0.0009 (0.0677) (0.0007) elec_proc 0.5781*** 0.0002 (0.0634) (0.0006) pol_plu 0.3755*** −0.0004 (0.0653) (0.0006) func_gov −0.9988*** 0.0006 (0.1098) (0.0010) civ_liberty −1.0861*** 0.0002 (0.0511) (0.0006) f_expression 0.1461** −0.0007 (0.0680) (0.0006) org_rights 0.8635*** 0.0013** (0.0682) (0.0006) rule_law −0.5654*** −0.0005 (0.0979) (0.0007) ind_rights 0.0868 −0.0003 (0.0899) (0.0006) elec −0.9556*** −0.0001 (0.2645) (0.0022) b_elec −0.4539* 0.0023 (0.2592) (0.0022) p_elec −0.6707** −0.0010 (0.2656 (0.0023) obs 1,014 793 1,014 793 1,012 614 502 614 502 612 Within R2 .5462 .6782 .5793 .6722 .5904 .0297 .0318 .0278 .0414 .0321 Between R2 .5802 .9177 .3887 .8362 .2891 .0563 .1596 .0560 .0776 .0472 Overall R2 .5410 .6778 .5472 .6672 .4782 .0680 .0275 .0671 .0414 .0675 F-test 3.93 2.28 3.01 3.46 3.88 4.51 3.72 4.49 3.41 4.51 (0.000) (0.016) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.001) (0.000) Islamic banks in democratic regime Islamic banks in autocratic regime

11 12 13 14 15 16 17 18 19 size 0.0149*** 0.0207 0.0103** 0.0156** 0.0140** 0.0016* 0.0004 0.0014 0.0003 (0.0054) (0.0074) (0.0051) (0.0067) (0.0058) (0.0009) (0.0009) (0.0009) (0.0009) liquidity 0.0333 0.0482* 0.0239 0.0430 0.0278 0.0115*** 0.0111*** 0.0110*** 0.0149*** (0.0220) (0.0255) (0.0223) (0.0256) (0.0219) (0.0035) (0.0039) (0.0035) (0.0039) (Continues) 18 ASUTAY AND MOHD SIDEK

TABLE 4 (Continued)

Islamic banks in democratic regime Islamic banks in autocratic regime

11 12 13 14 15 16 17 18 19 c 0.0856** 0.1461* 0.0660 0.2040** 0.1137 −0.0097 −0.0017 0.0106 0.0067 (0.0419) (0.0852) (0.0417) (0.0955) (0.0499) (0.0064) (0.0097) (0.0086) (0.0100) crisis −0.0247 −0.0149 −0.0183 (0.0190) (0.0169) (0.0199) pol_rights 0.0092** 0.0029** (0.0042) (0.0012) elec_proc 0.0119 0.0008 (0.0086) (0.0008) pol_plu 0.0004 −0.0004 (0.0027) (0.0005) func_gov −0.0081** −0.0006 (0.0039) (0.0012) civ_liberty 0.0026 −0.0003** (0.0027) (0.0001) f_expression −0.0010 −0.0010 (0.0033) (0.0006) org_rights 0.0010 0.0008 (−0.0041) (0.0009) rule_law 0.0011 −0.0004 (0.0041) (0.0007) ind_rights −0.0069 −0.0002 (0.0052) (0.0006) elec −0.0152** (0.0075) b_elec 0.0222*** (0.0079) p_elec 0.0020 (0.0074) obs 131 110 131 110 131 472 381 471 381 Within R2 .0856 .1058 .0548 .1017 .1299 .0364 .0414 .0341 .0552 Between R2 .0478 .1831 .0176 .1968 .0143 .9009 .5853 .0793 .1616 Overall R2 .0274 .1076 .0329 .1090 .0801 .0435 .0213 .0268 .0427 F-test 6.65 1.23 3.24 1.10 3.2 3.89 4.12 4.06 3.72 (0.000) (0.291) (0.000) (0.3738) (0.001) (0.000) (0.000) (0.000) (0.000)

Note: ***, ** and * represent 1%, 5% and 1% significance level. Robust standard errors are in parentheses for the coefficients. p-Values are in parentheses for F-statistics. depending on period, H2 is supported, as pre-election results, government effectiveness (g_eff), political stability dummy variable results in the expansion of loan in (p_stab), regulatory quality (reg) and voice and account- Islamic banks, but overall election variable has negative ability (va) have positive impact on loan growth whilst impact. This refers to mixed result for the H2. control of corruption (cor), the rule of law (rol) and ten- As the results reported in Table 5, all indicators of ure (ten) have negative impact on loan growth. Corrup- governance are significant in the overall sample provid- tion (cor) has negative impact in democratic regime but ing variable based support for H3, as in certain institu- positively affect loan growth in an autocratic regime. In tional and governance variables there is a positive democratic regimes where the market is almost fully impact; however, negative impact is produced in the case functioning, corruption disrupts its efficiency while in an of other governance variables. As can be seen in the autocratic regime where the market is not efficient; ASUTAY AND MOHD SIDEK 19

TABLE 5 Impact of governance on Governance loan growth (Dependent variable: Loan growth) Overall Islamic Democratic Autocratic

123 4 size 0.9386*** 0.0014 0.0317*** 0.0012 (0.0856) (0.0011) (0.0091) (0.0011) liquidity 0.0811*** 0.0139*** 0.0490* 0.0133*** (0.0206) (0.0043) (0.0257) (0.0041) c 10.6029*** 0.0431 0.2649*** 0.0018 (0.7454) (0.0161) (0.0882) (0.0100) crisis −1.1842 −0.0831*** (1.0055) (0.0265) cor −2.2403*** −0.0025** −0.0123** 0.0099* (0.3077) (0.0012) (0.0062) (0.0051) g_eff 1.2281*** 0.0113*** 0.0041 0.0093* (0.3226) (0.0039) (0.0250) (0.0054) p_stab 0.6255*** −0.0051* 0.0066 −0.0031 (0.2003) (0.0027) (0.0170) (0.0037) reg 1.8476*** −0.0029** −0.0134** −0.0072** (0.3434) (0.0012) (0.0065) (0.0034) rol −4.3364*** −0.0033** −0.0096* 0.0021 (0.4554) (0.0014) (0.0056) (0.0055) va 2.7031*** −0.0039* −0.0494** −0.0203*** (0.1778) (0.0039) (0.0238) (0.0058) ten −0.0417*** −0.0001 −0.0003 −0.0003** (0.0101) (0.0013) (0.0010) (0.0001) obs 912 523 116 351 Within R2 .6959 .0740 .1650 .921 Between R2 .6616 .0583 .2072 .5576 Overall R2 .6799 .0622 .0383 .0596 F-test 3.01 5.48 3.62 3.17 (0.001) (0.000) (0.001) (0.001)

Note: ***, ** and * represent 1%, 5% and 1% significance level. Robust standard errors are in parentheses for the coefficients. p-Values are in parentheses for F-statistics.

corruption expedites loan processes which leads to faster overall sample but negative in all sub-samples. This indi- approval and money being channelled into businesses. cates that freedom of expression, free media and freedom As can be seen, government effectiveness (g_eff) is the of association may undermine Islamic banking especially only significant in all Islamic banks and Islamic banks in with negative stigma attached to Islam and Muslims. autocratic regimes sub-sample. Regulatory quality (reg)is Muslim customers may opt for a loan in conventional positive in the overall sample but negative in all other banks with lower interest rates vis-a-vis loans from sub-samples. One possible explanation is that the regula- Islamic banks with a higher rate of return or bank char- tory quality for Islamic banks is still underdeveloped and ges. They can be divided into two distinct categories. The that the existing regulatory quality is not in favour of first category is the ones who patronize Islamic banking Islamic banks which undermines loan growth in Islamic as a tool for Shari'ah compliant transactions and other banks. The rule of law (rol) is negative and significant in businesses as part of their worship to God. Therefore, all samples except for the Islamic banks in autocratic regardless of the media portrays, they will stick to Islamic regimes sub-sample. These results are consistent with banking during good and adverse circumstances. The sec- earlier findings using Freedom House data. Furthermore, ond category consists of Muslims but those who do not voice and accountability (va) variable is positive in the practice Islam as a way of life, who is the opportunist 20 ASUTAY AND MOHD SIDEK who would take advantage of the best available rate are partly contributed by the larger sample of banks from regardless of whether the transaction is permissible or autocratic countries, which depicts that political rights as otherwise. Finally, longer tenure (ten) is found to be captured by the electoral process, political pluralism and harmful to loan growth. This indicates lack of innovative functionality of the government are all significant. It policies to capture the dynamic international and domes- should be noted that the functionality of the government tic markets, since the incumbent government may choose is negative and the other two criteria are positive. On a to remain faithful to old policies. similar note, the sub-variables for civil liberty are signifi- The results in Table 6 shows that four political risk cant with freedom of expression, organizational rights variables significantly affect loan growth, which provides and individual rights positively affecting loan growth. partial support for H4, as in the case of some political Having the rule of law significant and negative implies variables negative impact is located. As the results that more efforts to be expedited for law enforcement in depicts, democratic accountability (dem_acc) and law the majority of the sampled countries. The results are and order (lao) has positive effect on loan growth, while consistent with the findings in the previous section for government stability (g_stab) and internal conflict the overall sample. However, when the sample is (int_con) negatively affect loan growth for the overall restricted to Islamic banks only, all institutional variables sample. However, results for the sub-samples are mixed. except freedom of expression are no longer significant. Law and order (lao) is negative and significant in all sub- Lastly, Table 9 presents the results for the impact of samples reflecting either crime rate or lack of impartiality governance and political risks on loan growth. The signif- in the judiciary system affecting loan growth. In the case icance and signs of the coefficients mirror the results of a democratic regime, the absence of violence and other from the previous section. In summary, the results are political disorder, coupled with low socioeconomic pres- fairly robust in terms of significance and sign in the sures supports the growth of loan. As for autocratic majority of the regression. regimes, higher corruption, lack of impartiality in the judiciary system and high crime rate seem to suppress loan growth. 5.3 | The marginal impact of political regime, governance and political risks on loan growth 5.2 | Robustness check The results in Section 5.1 illustrate the direct impact of We run the overall sample and Islamic banks sample political variables on loan growth. In this section, we esti- using GLS and GMM for robustness check. In the case of mate interaction terms to locate the effect of loan growth feasible GLS, we report the estimation allowing the pres- change when the political variables change. In other ence of heteroscedasticity and AR(1) correction for auto- words, the marginal effect is approximated to examine correlation. For GMM, the instruments are limited to a how much loan growth is expected to increase or maximum of two-lags to avoid the problem of ‘too many decrease given changes in the political variables. It instruments’ (Roodman, 2009). In the case of Islamic should be noted that no conditional or multiplicative banks in democratic and autocratic countries, GMM is interactions are involved, which rules out the problems inefficient since the number of observations is of omitted terms, as it makes interpretations of marginal inadequate. effects invalid and the need for interpretations in terms As the results in Table 7 displays, generally, results of odd ratios, incidence ratios or hazard ratios. are reasonably consistent with the previous section Table 10 reports the summarized estimation results of regarding significance and signs. Bank-specific factors are the interaction variables testing for the short run and consistently positive albeit being insignificant in a few marginal impact of political variables on loan growth for regressions. For the institutional variables, polity2 is both the overall sample, Islamic banks only, Islamic banks in negative and significant with the disaggregated effect of the democratic regime and Islamic banks in autocratic being in a democratic regime (demo) is positive and sig- regimes. All regressions use the same set of control vari- nificant and negative for the autocratic regime (auto). ables as presented in Table A2 (see Appendix). The over- More specific disaggregation of institutions, however, are all results suggest that political stability plays a vital role not significant. towards loan growth. Pre-election year dummy variable The alternative variables to capture political regime has a significant impact on loan growth suggesting an which focuses on political rights and civil liberties are opportunistic behaviour of the ruling government toward both negative and significant which again indicate the elections in the following year. This reflects signs of eco- lack of these two criteria in our sample. Results in Table 8 nomic manipulation a year before elections under the ASUTAY AND MOHD SIDEK 21

TABLE 6 Impact of political risk Political risks on loan growth (Dependent variable: Loan growth) Overall Islamic Democratic Autocratic

567 8 size 0.7291*** 0.0021** 0.0243*** 0.0011 (0.1054) (0.0010) (0.0077) (0.0008) liquidity 0.0449* 0.0127*** 0.0312 0.0111*** (0.0247) (0.0038) (0.0230) (0.0035) c 12.4441*** 1.0343** 0.2484** 0.0053 (1.2219) (0.0138) (0.0948) (0.0074) crisis −2.0809** −0.0181* −0.0104 (0.9825) (0.0107) (0.0208) g_stab −7.6887*** −0.0086 −0.0008 −0.0042 (0.8841) (0.0074) (0.0272) (0.0109) socio −1.4471 0.0028 0.0468** 0.0106* (0.9257) (0.0079) (0.0229) (0.0055) cor2 −1.2734* 0.0036 0.0126 −0.0097** (0.6614) (0.0050) (0.0141) (0.0045) lao 2.5273*** −0.0117*** −0.0397* −0.0095** (0.7865) (0.0039) (0.0231) (0.0040) dem_acc 3.7970*** −0.0077* 0.0071 −0.0067 (0.5869) (0.0045) (0.0296) (0.0074) int_con −4.4506*** −0.0039 0.0687** 0.0025 (1.2792) (0.0121) (0.0295) (0.0132)

obs 981 605 136 409 Within R2 .5149 .0485 .1185 .0407 Between R2 .5028 .0325 .2854 .4729 Overall R2 .5108 .0824 .0231 .0545 F-test 3.32 4.55 3.39 3.56 (0.000) (0.000) (0.000) (0.000)

Note: ***, ** and * represent 1%, 5% and 1% significance level. Robust standard errors are in parentheses for the coefficients. p-Values are in parentheses for F-statistics.

democratic regime to increase popularity, as in the case elections. We argue that the Peacock-Wiseman (1961; see of Alesina (1988). Based on Nordhaus' (1975) political also, Henrekson, 1990) hypothesis of displacement effect business cycle model, the opportunistic incumbent gov- of tax-government expenditure nexus may apply in our ernment implements an expansionary monetary policy to context of loan growth pre- and post-election periods in temporarily expand economic activities, which is con- the sense that pre-election loan growth replaces the natu- ducted by increasing bank loans to government, busi- ral trend by taking the pace of loan growth to a higher nesses and individuals leading to intensification of loan level. disbursement or ‘easy’ loans leading to higher loan The evidence presented on political risks is rather growth. In the overall sample, all the election, pre-elec- clear; as the results suggest lower political risks promote tion and post-election dummy variables are significant loan growth. The loan growth-government stability inter- implying displacement effect. While the results relating action suggests government unity, legislative strength to pre-election year indicate growth in loans due to gov- and popular support smooths government's ability to ful- ernment's manipulation of the economy or artificially fil their declared manifestoes and policies. Stability also boosting the economy, government, businesses and the means the incumbent government has the ability to win general public may become accustomed to such ‘easy’ the election to stay in the office after elections, which loans and the trend may continue even after the enables them to carry out any long-term policies. The 22 ASUTAY AND MOHD SIDEK

TABLE 7 Impact of political regime on loan growth (Overall sample: Conventional and Islamic banks, Dependent variable: Loan growth)

Overall sample (Islamic and Conventional banks)

12345

Regression GLS GMM GLS GMM GLS GMM GLS GMM GLS GMM size 0.7215*** 0.6523*** 0.8373*** 0.6525*** 0.8542*** 0.6764*** 0.9022*** 0.7598*** 0.7185*** 0.7251*** (0.2658) (0.2407) (0.3271) (0.2315) (0.2839) (0.2254) (0.3246) (0.2354) (0.2699) (0.1987) liquidity 0.1107*** 0.1003** 0.1651*** 0.1005*** 0.1361*** 0.1102*** 0.1655*** 0.1565*** 0.1956*** 0.1675** (0.0356) (0.0431) (0.0224) (0.0412) (0.0456) (0.0374) (0.0573) (0.0587) (0.0736)) (0.0684) constant 2.7420** 3.7532** 2.3672*** 3.6544** 2.5983*** 3.9456** 3.1006*** 3.7954** 3.2119*** 2.7590** (1.2352) (1.5320) (0.6268) (1.5164) (0.8960) (1.5422) (0.8449) (1.5479) (0.9991) (1.0820) crisis −1.2854*** −1.1255** −1.2123*** −1.1226** −1.3009** −1.2039** −1.5310** −1.8512** −1.6544** −2.3225** (0.4709) (0.4831) (0.3898) (0.4769) (0.5310) (0.4910) (0.5912) (0.7090) (0.6747) (0.9633) polity2 −3.0025*** −3.5219*** (0.8831) (0.9655) demo 0.4215*** 0.3954*** 0.5528*** 0.4527*** 0.5003** 0.4125** (0.01500) (0.1373) (0.1707) (0.1286) (0.2050) (0.1755) autoc −0.2293*** −0.2543*** (0.0715) (0.0792) xrreg −0.6053*** −0.5428** (0.1615) (0.2162) xrcomp −0.0101 −0.0052 (0.0237) (0.2004) xopen 0.0127 0.0053 (0.1090) (0.0112) xconst −0.2192 −0.2009 (0.1768) (0.1925) parreg −0.1985 −0.2500 (0.1811) (0.3215) parcomp 0.0365 0.0032 (0.1251) (0.0134) exrec 0.0062 0.0425 (0.0751) (0.0815) exconst −0.0325 −0.0521 (0.0639) (0.0625) polcomp −0.0013 −0.0015 (0.0542) (0.0632) Observation 1,026 296 1,026 296 1,026 286 1,026 262 796 222 Wald Chi2 Stat 0.00 0.00 0.00 0.00 0.00 Hansen J-test 0.596 0.599 0.574 0.523 0.498 Islamic banks

678910

Regression GLS GMM GLS GMM GLS GMM GLS GMM GLS GMM size 0.0054** 0.0096** 0.0068** 0.0105** 0.0235** 0.0266** 0.0119** 0.0291** 0.0155** 0.0253** (0.0022) (0.0038) (0.0030) (0.0042) (0.0093) (0.0105) (0.0047) (0.0114) (0.0061) (0.0099) liquidity 0.0209** 0.0532** 0.0309** 0.0352** 0.0323** 0.0453** 0.0543** 0.0498** 0.0568** 0.0512** (0.0083) (0.0220) (0.0132) (0.0150) (0.0124) (0.0195) (0.0213) (0.0207) (0.0232) (0.0219) constant 0.01056 0.1024 0.0254 0.1024 0.1324 0.2015 0.1566 0.0758 0.1059 0.1504 ASUTAY AND MOHD SIDEK 23

TABLE 7 (Continued)

Islamic banks

678910

Regression GLS GMM GLS GMM GLS GMM GLS GMM GLS GMM (0.0210) (0.2543) (0.0953) (0.2353) (0.3215) (0.2354) (0.2017) (0.0973) (0.2015) (0.0932) crisis −0.0325** −0.0452** −0.0542** −0.0053 −0.0219** −0.0095 −0.0498** −0.0056 −0.0512** −0.0048 (0.0133) (0.0193) (0.0208) (0.0102) (0.0099) (0.0106) (0.0214) (0.0152) (0.0202) (0.0132) polity2 0.0124** 0.0254** (0.0049) (0.0100) demo 0.0023 0.0036 0.0012 0.0014 0.0017 0.0015 (0.0035) (0.0105) (0.0026) (0.0152) (0.0155) autoc −0.0035** −0.0051** (0.0016) (0.0021) xrreg −0.0032** −0.0006 −0.0033** −0.0001 (0.0014) (0.0052) (0.0014) (0.0051 xrcomp 0.0011 0.0033 (0.0045) (0.0102) xopen 0.0004 0.0002 (0.0012) (0.0036) xconst −0.0008 −0.0011 −0.0022 −0.0012 (0.0013) (0.0016) (0.0031) (0.0017) parreg −0.0010 −0.0012 −0.0019 −0.0015 (0.0027) (0.0056) (0.0018) (0.0053) parcomp 0.0002 0.0003 0.0003 0.0004 (0.0032) (0.0045) (0.0035) (0.0041) exrec exconst polcomp Observation 626 178 626 178 626 178 626 95 604 107 Wald Chi2 Stat 0.00 0.00 0.00 0.00 0.00 Hansen J-test 0.311 0.313 0.313 0.122 0.129

interaction of socioeconomic variable and loan growth wanting to comfort in life (such as personal loans, car indicates lower unemployment, higher consumer confi- loans, housing loans). The ability of the government dence and lower poverty reduces risks of socioeconomic reacting to this requirement signals government respon- pressure and consequently promotes growth. The siveness through public to the public. absence of internal conflict in the form of civil war, civil disorder or any kind of political violence promotes political stability which provides a favourable environment for 6 | CONCLUSION banking related activities. Democratic accountability reflects lower risks for democratic countries and higher This article explored the role of political regime and insti- risks for autocratic countries. In this case, democratic tutions, governance and political risks on Islamic bank- accountability evaluates the responsiveness of the gov- ing growth, proxied by loan growth across both ernment towards public demand especially financially, as conventional and Islamic banks in Muslim dominated loans are essential for business expansion and individuals countries. Our earlier hypothesis is that there exists a 24 ASUTAY AND MOHD SIDEK

TABLE 8 Impact of political regime on loan growth (Overall sample: Conventional and Islamic banks, Dependent variable: Loan growth)

Overall sample (Islamic and Conventional banks)

12 3 45

Regression GLS GMM GLS GMM GLS GMM GLS GMM GLS GMM size 0.5473*** 0.6734** 0.5116*** 0.6642*** 0.7901*** 0.6879** 0.6799*** 0.5491** 0.5872*** 0.6176** (0.1511) (0.2682) (0.1717) (0.2363) (0.2431) (0.2635) (0.2346) (0.2236) (0.2099) (0.2506) liquidity 0.3594*** 0.4561** 0.3870*** 0.3866*** 0.4532*** 0.4099*** 0.3225*** 0.4132*** 0.2861*** 0.4328*** (0.1012)) (0.1801) (0.1423) (0.1243) (0.1564) (0.1497) (0.9988) (0.1488) (0.0869) (0.1219) constant 7.3471*** 6.0932*** 5.4416*** 5.2768*** 7.3287*** 6.1975*** 4.7511*** 5.7429*** 5.7683*** 4.8761*** (2.7415) (2.1418) (1.8895) (1.4823) (1.6247) (1.6851) (1.5844) (1.6315) (1.9234) (1.4976) crisis −1.2017 −1.1322 −1.0562 −1.3252 −1.5471 0.9976 −1.0017 −0.9961 −0.9329 −1.0245 (1.1536) (1.5473) (1.1254) (1.4278) (1.3211) (1.3767) (1.4132) (1.3872) (0.8763) (1.3481) demo 0.5497*** 0.6753** (0.1712) (0.2773) pol_rights −1.4875*** 1.6785*** (0.5109) (0.5396) elec_proc 0.6859*** 0.7756** (0.2136) (0.3301) pol_plu 0.4686*** 0.4975*** (0.1505) (0.1360) func_gov −0.5891*** −0.5900** (0.1971) (0.2331) civ_liberty −2.1197*** −1.9789*** (0.8106) (0.5416) f_expression 0.1132** 0.2675** (0.0479) (0.1049) org_rights 0.7654*** 0.6423** (0.2460) (0.2513) rule_law −0.4532*** −0.0287 (0.1626) (0.3861) ind_rights 0.0098 0.0079 (0.0108) (0.0207) elec −1.2527*** −1.0894** (0.4142) (0.4256) b_elec −0.0564 0.0087 (0.2119) (0.1776) p_elec −0.0869 0.0782 (0.1865) (0.1968) Observation 1,028 297 807 182 1,028 297 807 147 1,026 274 Wald Chi2 Stat 0.00 0.00 0. 0.00 0.00 0.00 Hansen J-test 0.545 0.336 0.549 0.226 0.478 Islamic banks

678910

Regression GLS GMM GLS GMM GLS GMM GLS GMM GLS GMM size 0.0054 0.0064 0.0003 0.0001 0.0008 0.0003 0.0003 0.0001 0.0007 0.0001 (0.0049) (0.0079) (0.0008) (0.0057) (0.0010) (0.0065) (0.0009) (0.0055) (0.0009) (0.0054) liquidity 0.0322** 0.0448** 0.0115** 0.0341** 0.0113** 0.0421** 0.0125** 0.0132** 0.0133** 0.0114** ASUTAY AND MOHD SIDEK 25

TABLE 8 (Continued)

Islamic banks

678910

Regression GLS GMM GLS GMM GLS GMM GLS GMM GLS GMM (0.0137) (0.0183) (0.0046) (0.0145) (0.0049) (0.0166) (0.0053) (0.0052) (0.0051) (0.0045) constant 0.0075 0.0074 −0.0053 −0.0021 0.0065 0.0076 0.0054 0.0032 0.0057 0.0031 (0.0087) (0.0073) (0.0059) (0.0071) (0.0077) (0.0077) (0.0083) (0.0073) (0.0085) (0.0071) crisis −0.0197 −0.0055 −0.0101 −0.0098 −0.0077 −0.0088 −0.0065 −0.0015 −0.0100 −0.0087 (0.0201) (0.0197) (0.0213) (0.0187) (0.0211) (0.0176) (0.0227) (0.0154) (0.0097) (0.0097) demo 0.0005 0.0003 (0.0006) (0.0011) pol_rights 0.0013 0.0009 (0.0053) (0.0043) elec_proc 0.0005 0.0010 (0.0007) (0.0013) pol_plu −0.0003 −0.0004 (0.0005) (0.0007) func_gov 0.0007 0.0011 (0.0011) (0.0087) civ_liberty 0.0007 0.0003 (0.0011) (0.0045) f_expression −0.0003 −0.0001 (0.0006) (0.0011) org_rights 0.0142** 0.0098** (0.0056) (0.0040) rule_law −0.0004 −0.0002 (0.0007) (0.0010) ind_rights −0.0002 −0.0002 (0.0005) (0.0013) elec −0.0001 −0.0002 (0.0023) (0.0035) b_elec 0.0027 0.0033 (0.0033) (0.0101) p_elec −0.0011 0.0005 (0.0035) (0.0056) Observation 628 176 516 102 628 176 516 94 626 115 Wald Chi2 Stat 0.00 0.00 0.00 0.00 0.00 Hansen J-test 0.316 0.287 0.325 0.137 0.299

Note: Standard errors are in parentheses. *** and ** denote statistical significance at 1% and 5%. The H0 for Chi-squared test is that all the coefficients except the constant are jointly equal to zero. The Hansen J-test of overidentifying restrictions test the H0 that the instruments are valid. relationship between loan growth and political institu- political business cycle is evident in our results for demo- tions where good institutions promote the development cratic countries where the current ruling government of Islamic banking in terms of higher loan growth. artificially boost the economy to secure votes in the next Results suggest that loan growth is higher in demo- elections. Interestingly, the effect of corruption differs in cratic regimes as compared to autocratic regimes, democratic and autocratic regime; as corruption in demo- although more than two-thirds of the total assets in cratic countries negatively affect loan growth, but the Islamic banking are within the autocratic regime-based effect is positive in autocratic countries implying different countries mainly in the GCC region. The opportunistic regime framework leading to different response and 26 ASUTAY AND MOHD SIDEK

TABLE 9 Impact of governance and political risks on loan growth (Dependent variable: Loan growth)

Governance Political risks

Overall Islamic Overall Islamic

12 1 2

Regression GLS GMM GLS GMM GLS GMM GLS GMM size 1.2451*** 1.0243** 0.0073*** 0.0083*** Total assets 0.9544*** 1.0135*** 0.0045** 0.0032** (0.4717) (0.5122) (0.0025) (0.0029) (0.3563) (0.3099) (0.0022) (0.0015) liquidity 0.0981** 0.0764 −0.0014 −0.0011 Total loans 0.1103*** 0.1008*** 0.0221*** 0.0256** (0.0491) (0.0066) (0.0010) (0.0247) (0.0418) (0.0351) (0.0076) (0.0117) constant 9.4211*** 7.9851 0.0544 0.0643 Constant 9.4358** 11.5421** 2.6341** 2.8250** (3.5686) (3.002) (0.0535) (0.0585) (4.6945) (5.2342) (1.1306) (1.2177) crisis −1.3854 −0.9453 −0.0082 −0.0069 crisis −2.3564** −2.1056** −0.0200** −0.0191 (1.0211) (1.0052) (0.0073) (0.0071) (1.1221) (0.9031) (0.0091) (0.0245) corruption −1.2454*** −0.9785 −0.0049*** −0.0032 g_stab −8.1569*** −8.8871** −0.0089 −0.0095 (0.4681) (1.1022) (0.0011) (0.0029) (3.0890) (4.361) (0.0078) (0.0101) g_eff 1.0125*** 1.3264 0.0253*** 0.0925 socio −1.2532 −1.0953 0.0051 0.0045(0.0039) (0.3642) (1.1255) (0.0047) (0.1005) (0.9965) (0.9556) (0.0043) p_stab 0.0258 0.7581 −0.0072 0.0043 corruption2 −1.3149** −1.3849** 0.0062 0.0051 (0.2559) (0.6953) (0.0064) (0.0050) (0.5717) (0.5412) (0.0053) (0.0060) reg 2.8546*** 2.3157** −0.0058** −0.0032 lao 1.9987*** 2.0053** −0.0287*** −0.0252** (0.9581) (1.1579) (0.0013) (0.0107) (0.6663) (0.9511) (0.0103) (0.0120) rol −5.3654*** −3.1151 −0.0065*** −0.0054 dem_acc 3.9921*** 4.2695*** −0.0080 −0.0075 (1.7885) (2.5975) (0.0012) (0.0105) (1.1406) (1.6112) (0.0065) (0.0070) va 2.9577 2.3541 −0.0077 0.0063 int_con −3.8125*** −3.2654*** −0.0036 −0.0025 (1.6573) (1.8542) (0.0059) (0.0114) (1.1914) (1.0035) (0.0123) (0.0101) ten −0.0113 0.0344 −0.0001 −0.0001 (0.1652) (0.1568) (0.0011) (0.0045) Observation 922 256 535 140 973 276 599 166 Wald Chi2 Stat 0.00 0.00 0.00 0.00 Hansen J-test 0.345 0.202 0.415 0.211

outcomes, as such a relation in autocratic regimes might The results cannot, however, be generalized for over- be the result of patronage or clientelism based political all Islamic banking due to several possible caveats. First, economy. In addition, the case of countries under autoc- anecdotal evidence suggests that the degree and type of racy, corruption speed up business processes by removing democracy and autocracy vary across the sampled coun- bureaucratic impediments and speeding up the processes. tries: Turkey, Bangladesh and Pakistan is a parliamentary Results also suggest that certain political-institutional representative democratic republic, Malaysia is a federal framework is designed for conventional banks, therefore, representative democratic constitutional monarchy and may not necessarily fit into the Islamic banking require- Indonesia is a presidential representative democratic ments. The findings also suggest that different regimes republic. Saudi Arabia, Qatar and Oman reign under an are affected by different forms of political risks vis-à-vis absolute monarchy, while Bahrain, Kuwait, Jordan and loan growth. The results show that effective socioeco- to some extent, the UAE is under constitutional monar- nomic conditions and low internal conflict provide a chy system. Yemen, Egypt, Iran and Sudan practice presi- favourable environment for loan growth in democratic dential democracy, but the actual practices are akin to countries, while in autocratic regimes, higher corruption those in authoritarian regime type. To alleviate this prob- and lax enforcement of law amount lower loan growth. lem, we categorized the countries not according to the de ASUTAY AND MOHD SIDEK 27

TABLE 10 Summary of interaction variables (loan growth × political variables; Dependent variable: Loan growth)

Overall sample Islamic banks Democratic regime Autocratic regime Political regime polity2 0.1303*** −0.0529*** (0.0027) (0.0064) auto 0.1432*** (0.0034) demo 0.2408*** (0.0048) elec 0.7765*** 0.5898*** −0.1307 (0.1335) (01558) (0.5009) b_elec 0.8230*** 0.9232*** 0.9652*** (0.1108) (0.0429) (0.0211) p_elec 0.9628*** 0.6150*** 0.2149 (0.0351) (0.1397) (0.7450) pr 0.2399***(0.0012) 0.1780*** 0.1617*** (0.0017) (0.0011) cl 0.2589*** 0.1765*** 0.1602*** (0.0027) (0.0019) (0.0012) Governance cor −0.8444*** −0.2987*** 1.9587*** −0.4942*** (0.0237) (0.0521) (0.1959) (0.0426) g_eff −0.7980*** 0.2193*** 0.7501*** −0.5372*** (0.0637) (0.0397) (0.0228) (0.0488) p_stab −0.4655*** −0.3903*** 0.4208** −0.4338*** (0.0092) (0.0282) (0.1752) (0.0212) reg −0.7811*** −0.0960** 1.6198*** −0.4113*** (0.0614) (0.0441) (0.0732) (0.0371) rol −0.9791*** −0.1279*** 1.5311*** −0.4285*** (0.0284) (0.0467) (0.0917) (0.0399) va −1.2149*** −0.8101*** −1.8511*** −0.7136*** (0.0294) (0.0187) (0.0339) (0.0123) ten 0.0481** 0.0678*** 0.1914** 0.0592** (0.0032) (0.0019) (0.0047) (0.0015) Political risks g_stab 1.4193*** 1.2965*** 1.2331*** 1.2975*** (0.0049) (0.0089) (0.0089) (0.0121) socio 2.3964*** 1.4258*** 1.2081*** 1.7309*** (0.0469) (0.0274) (0.0161) (0.0019) int_con 1.8681*** 1.2779*** 1.1518*** 1.3811*** (0.0181) (0.0110) (0.0105) (0.0149) cor2 2.3391*** 1.2222*** 1.0587*** 1.5998*** (0.0273) (0.0275) (0.0129) (0.0498) dem_acc 1.4565*** 1.4161*** 1.3378*** 1.4813*** (0.0073) (0.0185) (0.0080) (0.0283) lao 2.0542*** 1.6814*** 2.8899*** 1.5034*** (0.0231) (0.0441) (0.0600) (0.0072)

Note: Full results are available upon request. ***and ** represent 1% and 5% significance level. 28 ASUTAY AND MOHD SIDEK jure classification but based on the de facto definition. capturing about 25% of the total financial system in the Countries with a score of more than five from the Polity country. Similarly, the UK governments in general and IV database are classified as democratic while countries the Labour governments in the first decade of the new with scores of less than five are classified as autocratic. century in particular facilitated the development of The cut-off score is based on Eichler (2014). Second, the Islamic finance in the UK with the objective of making composition of the loan, for example, loan or financing London an important Islamic finance hub, while the lack based on qardhassan, murabahah, mudharabah, of political will in Turkey until 2013, despite the Islamic musharakah, salam or istisna4 may also differ amongst roots of AKP government, had not provided the necessary the Islamic banks. Such differences are due to differences impetus for the development of Islamic finance and its in Shari'ah interpretation. In addition, the definition of financial architecture in Turkey, as Turkey has only the loan may differ from one country to another. Third, recently become committed to developing its Islamic given the different nature of conventional and Islamic finance industry by generating a new model of Islamic banks, loans may not be entirely comparable across bank banking industry development. Indonesia's sluggish types even within the same country. Karim (2011) dis- development of Islamic finance industry can be explained cusses the varying accounting practices across countries in a similar vein; as until the regime change in the 1990s, with Islamic banking suggesting that certain balance the opportunity space for Islamic financial expansion was sheet and income statement items may not be entirely not available; and therefore, they opted for a civil society comparable across bank types even within the same based institutional development. However, with the country. We have endeavoured to circumvent this prob- democratization of the country, the expansion of Islamic lem by relying on and standardizing data extraction from banking and finance and the necessary regulative archi- the same source as much as possible, for example, tecture has gained new momentum. As these examples Bloomberg and IRTI. indicate, political environment plays an essential role in An important policy implication is that the develop- the diffusion, development and operation of Islamic ment of Islamic banking is determined very much by the finance. political will of the ruling government or the rulers. Regardless of being democratic or autocratic, Islamic ORCID banking flourishes by given political support of the Mehmet Asutay https://orcid.org/0000-0003-4939-6053 authorities which expedite decision making, creating specific sets of laws and regulation for Islamic banking, ENDNOTES research and development in Islamic banking, Islamic 1 While the right terminology should be ‘financing’ for Islamic banking friendly policies and regulations as well as capi- banks due to the intended meaning suggested by the value system tal for Islamic banking; all of which are lifeline for the of Islamic moral economy, in order to ensure consistency with the survival of Islamic banking. Islamic banks still have to existing literature as well as making a coherent comparison with ‘ ’ develop in term of competitiveness, maturity, sophistica- conventional banks, loans is used as a term for Islamic banks in this article. tion, management, a technological adaptation which may 2 explain the reason for the differences in behaviour Murabahah: Sale with mark-up; Tawarruq: Short-term financing based on buying commodity through deferred payment and towards political variables as identified by this study. immediate sale to a third party on the spot (for further details see: Islamic banking alongside with conventional banking Ayub, 2007; Asutay, 2015). has subsequently embraced technologies, products, mar- 3 Results are not reported to conserve space but available upon keting strategies, policies, regulations; and these seem to request. have converged over time. Hence, this infers that political 4 Murabahah: Sale with mark-up; Mudarabah: Profit sharing; regime is not only a determinant of Islamic banking pro- Qardhassan: Benevolent loan; Musharakah: Partnership; Salam: gress, preferably, the political commitment of the ruling Payment for deferred delivery of goods; Istisna': Payment made in government is vital to set forth a stable and robust stages according to work progress; Tawarruq: Short-term financ- Islamic banking standing. As long as the ruling governing based on commodity buy and immediate sale; Sukuk: Islamic ment or authority is willing to embrace financial-friendly bonds (for further details see: Ayub, 2007, Asutay, 2015). policy framework, Islamic banking will be able to flour- ish anywhere implying the essential nature of the political process as an essential determinant. This is visible in DATA AVAILABILITY STATEMENT the everyday practice of Islamic finance, as the Malaysian The data that support the findings of this study are avail- government's proactive nature in developing industry has able from the corresponding author upon reasonable resulted in Islamic banks and financial institutions request. ASUTAY AND MOHD SIDEK 29

ORCID Barro, R. J., & McCleary, R. M. (2003). Religion and economic Mehmet Asutay https://orcid.org/0000-0003-4939-6053 growth. NBER Working Paper No: 9682. Cambridge, MA: National Bureau of Economic Research. Baum, C. F., Caglayan, M., & Talavera, O. (2010). Parliamentary REFERENCES election cycles and the Turkish banking sector. Journal of Acemoglu, D. (2008). Oligarchic versus Democratic Societies. Jour- Banking & Finance, 34, 2709–2719. – nal of the European Economic Association, 6(1), 1 44. Baum, C. F., Caglayan, M., Schafer, D., & Talavera, O. (2008). Polit- Acemoglu, D., Johnson, S., & Robinson, J. A. (2005). Institutions as ical patronage in Ukraine banking. The Economics of Transi- the fundamental cause of long-run economic growth. In P. tion, 16(3), 537–557. Agion & S. Durlauf (Eds.), Handbook of economic growth (pp. Baum, M., & Lake, D. (2003). The political economy of growth: – 385 472). Amsterdam: North Holland. Democracy and human capital. American Journal of Political Aglietta, M. (1976). A theory of capitalist regulation: The US experi- Science, 47(2), 333–347. ence. London: Verso. Beck, T., Demirguc-Kunt, A., & Levine, R. (2003). Law, endow- Aidt, T. S., Dutta, J., & Sena, V. (2008). Governance regimes, cor- ment, and finance. Journal of Financial Economics, 70(2), ruption and growth: theory and evidence. Journal of Compara- 137–181. – tive Economics, 36, 195 220. Beck, T., Demirguc-Kunt, A., & Merrouche, O. (2013). Islamic vs. Aisen, A., & Jose Veiga, F. (2013). How does political instability conventional banking: Business model, efficiency and stability. affect economic growth? European Journal of Political Econ- Journal of Banking & Finance, 37, 433–447. – omy, 29, 151 167. Bitar, M., Hassan, M. K., & Walker, T. (2017). Political systems and Alberto, Alesina (1988). Macroeconomics and Politics: NBER Macro- the financial soundness of Islamic banks. Journal of Financial economics Annual, Cambridge: Cambridge University Press. Stability, 31,18–44. Alesina, A., & Perotti, R. (1996). Income distribution, political instabil- Bjerksund, P., & Ekern, S. (1990). Managing investment opportuni- – ity and investment. European Economic Review, 40,1203 1228. ties under price uncertainty: From ‘last chance’ to ‘wait and Alesina, A., Ozler, S., Roubini, N., & Swagel, P. (1996). Political see’ strategies. Financial Management, 19(3), 65–83. instability and economic growth. Journal of Economic Growth, Blundell, R., & Bond, S. (1998). Initial conditions and moment – 1, 189 211. restrictions in dynamic panel data models. Journal of Econo- Alesina, A., Roubini, N., & Cohen, G. D. (1997). Political cycles and metrics, 87, 115–143. the macroeconomy. Cambridge, MA: The MIT Press. Bongini, P., Laeven, L., & Majnoni, G. (2002). How good is the Arab Times. (2015). Commercial banks to provide Shari'ah compli- market at assessing bank fragility? A horse race between ant products and services. different indicators. Journal of Banking & Finance, 26(5), 1011– Ashraf, B. N. (2017). Political institutions and bank risk-taking 1028. – behaviour. Journal of Financial Stability, 29,13 35. Borooah, V. K. (1985). The interaction between economic policy Asutay, M. (2007a). A political economy approach to islamic eco- and political performance. In R. C. O. Matthews (Ed.), Economy nomics: Systemic understanding for an alternative economic and democracy. Houndmills: Macmillan. – system. Kyoto Bulletin of Islamic Area Studies, 1(2), 3 18. Boyer, R. (1990). The regulation school: A critical introduction. New Asutay, M. (2007b). Conceptualisation of the second best solution York: Columbia University Press. in overcoming the social failure of islamic finance: Examining Boyer, R., & Saillard, Y. (Eds.). (2002). Regulation theory: The state the overpowering of homoislamicus by homoeconomicus. IIUM of the art. New York: Routledge. – Journal of Economics and Management, 15(2), 167 195. Butkiewicz, J. L., & Yanikkaya, H. (2006). Institutional quality and Asutay, M. (2012). Conceptualising and locating the social failure of economic growth: Maintenance of the rule of law or democratic islamic finance: Aspirations of Islamic moral economy vs. the institutions, or both. Economic Modelling, 23, 648–661. realities of islamic finance. Journal of Asian and African Stud- Caporaso, J. A., & Levine, D. P. (1992). Theories of political econ- – ies, 11(2), 93 113. omy. New York: Cambridge University Press. Asutay, M. (2013). Islamic moral economy as the foundation of Cheibub, J., Ghandi, J., & Vreeland, J. (2010). Democracy and dicta- islamic finance. In V. Cattelan (Ed.), Islamic finance in Europe: torship revisited. Public Choice, 143(1), 67–101. Towards a plural financial system. Cheltenham: Edward Elgar. Chen, P.-F., & Liu, P.-C. (2013). Bank ownership and performance, Asutay, M. (2015). Islamic finance: Principles, institutions and and the politics: Evidence from Taiwan. Economic Modelling, – development. Trade and Finance, 2015,4 26. 31, 578–585. Ayub, M. (2007). Understanding Islamic finance. Chichester, West Clague, C., Keefer, P., Knack, S., & Olson, M. (1996). Property and Sussex: John Wiley & Sons Ltd. contract rights in autocracies and democracies. Journal of Eco- The Banker, (2013). Top Islamic Financial Institutions-Special nomic Growth, 1(2), 243–276. Report, London: Financial Times. Cole, S. (2009). Fixing market failures or fixing elections? Agricul- The Banker, (2014). Top Islamic Financial Institutions-Special tural credit in India. American Economic Journal: Applied Eco- Report, London: Financial Times. nomics, 1(1), 219–250. The Banker, (2017). Top Islamic Financial Institutions-Special Collins, Kathleen (2006). Clan Politics and Regime Transition in Report, London: Financial Times. Central Asia, Cambridge : Cambridge University Press. Barro, R. J. (1989). A cross-country study of growth, saving and gov- Darby, J., Li, C., & Muscatelli, A. (2004). Political uncertainty, pub- ernment. NBER Working Paper No. 2855. Cambridge, MA: lic expenditure and growth. European Journal of Political Econ- National Bureau of Economic Research. omy, 20, 153–179. 30 ASUTAY AND MOHD SIDEK

De Long, J. B., & Shleifer, A. (1993). The princes and merchant: Helliwell, J.F. (1992). Empirical linkages between democracy and European city growth before industrial revolution. The Journal economic growth. NBER Working Paper No. 4066. Cambridge: of Law & Economics, 36(2), 671–702. National Bureau of Economic Research. Debs, A., & Goemans, H. E. (2010). Regime type, the fate of leaders Henrekson, M. (1990). The Peacock and Wiseman displacement and war. American Political Science Review, 104(3), 430–445. effect. A reappraisal and a new test. European Journal of Politi- Demetrescu, M., & Hanck, C. (2012). A simple nonstationary-vola- cal Economy, 6(2), 245–226. tility robust panel unit root test. Economics Letters, 117,10–13. Hernandez, D., & Vadlamannati, K. C. (2016). Politics of religiously Devereux, M., & Wen, J. F. (1998). Political instability, capital taxa- motivated lending: An empirical analysis of aid allocation by tion and growth. European Economic Review, 42, 1635–1651. the Islamic Development Bank. Journal of Comparative Eco- Dick, W. G. (1974). Authoritarian versus nonauthoritarian nomics, 21,1–20. approaches to economic development. Journal of Political Econ- Herwartz, H., & Siedenburg, F. (2008). Homogeneous panel unit omy, 82(4), 817–827. root test under cross-sectional dependence: finite sample modi- Dinc, L. S. (2005). Politicians and banks: Political influences on gov- fications and wild bootstraps. Computational Statistics and ernment owned banks in emerging markets. Journal of Finan- Data Analysis, 53, 137–150. cial Economics, 77(2), 453–479. Huang, Y. (2010). Political Institutions and Financial Development. Eichler, S. (2014). The political determinants of sovereign bond World Development, 38(12), 1667–1677. yield spread. Journal of International Money and Finance, 46, Huntington, S. P., & Dominguez, J. I. (1975). Political development. 82–103. In F. I. Greenstein & N. W. Polsby (Eds.), Handbook of political Eichler, S., & Sobanski, K. (2016). National politics and bank science (pp. 1–114). Reading: Addison-Wesley. default risk in the Eurozone. Journal of Financial Stability, 26, IFSB. (2019). Islamic Financial Services Industry Stability Report 247–256. 2019. Kuala Lumpur: IFSB. Elashker, A., & Wilson, R. (2006). Islamic economics: A short history. Iyer, S. (2010). Religion and economic development. In S. N. Leiden: Brill. Durlauf & L. E. Blume (Eds.), Economic growth (The New Pal- Ernst & Young, (2009). Ernst & Young's Islamic Funds & Investment grave Economics Collection). London: Palgrave Macmillan. Report 2009, London: Ernst & Young. Jackowicz, K., Kowalewski, O., & Kozlowski, L. (2013). The influ- Ernst & Young. (2016). World Islamic Banking Competitiveness ence of political factors on commercial banks in Central Euro- Report 2016. pean countries. Journal of Financial Stability, 9(4), 759–777. Feng, Y. (1997). Political stability and economic growth. British Jetter, M. (2014). Volatility and growth: Governments are key. Journal of Political Science, 27(3), 391–418. European Journal of Political Economy, 36,71–88. Flachaire, E., Garcia-Penalosa, C., & Konte, M. (2014). Political ver- Jones, B. F., & Olken, B. A. (2008). The anatomy of start-stop sus economic institutions in the growth process. Journal of growth. The Review of Economics and Statistics, 90(3), 582–587. Comparative Economics, 42, 212–229. Jong-a-Pin, R. (2009). On measurement of political instability and Fraser, D. R., Zhang, H., & Derashid, C. (2006). Capital structure its impact on economic growth. European Journal of Political and political patronage: The case of Malaysia. Journal of Bank- Economy, 25,15–29. ing & Finance, 30(4), 1291–1308. Karim, R. A. A. (2011). International accounting harmonization, Freeman, J. R., Hays, J. C., & Stix, H. (2000). Democracy and mar- banking regulation and Islamic banks. The International Jour- kets: The case of exchange rate. American Journal of Political nal of Accounting, 36, 169–193. Science, 44(3), 449–468. Khwaja, A. I., & Mian, A. (2005). Do lenders prefer politically con- Frey, B. S. (1978). Modern political economy. Oxford: Martin nected firms? Rent provision in an emerging financial market. Robertson. The Quarterly Journal of Economics., 120(4), 1371–1411. Frohlich, Norman & Oppenheimer, Joe A. (1978). Modern Political Kleymenova, A., Rose, A. K., & Wieldek, T. (2016). Does govern- Economy, New York: Prentice-Hall. ment intervention affect banking globalization? Journal of the Gabgub, Aburawi Issa (2009). Analysis of Non-Performing Loans in Japanese and International Economies, 42, 146–161. the Libyan State-Owned Commercial Banks: Perception Analysis Kohli, A. (1986). Democracy and development in Lewis. In J. P. of the Reasons and Potential Methods for Treatment, PhD The- Lewis & V. Kallab (Eds.), Development strategies reconsidered sis, School of Government and International Affairs, Durham: (pp. 153–182). New Brunswick: Transaction Books. Durham University. Kuran, T. (2004). Why the middle east is economically underdevel- Ghosh, S. (2015). Political transition and bank performance. Jour- oped: Historical mechanism of institutional stagnation. Journal nal of Comparative Economics, 44(2), 372–382. of Economic Perspective, 18(3), 71–90. Gimma, S., & Shortland, A. (2008). The political economy of finan- Kuran, T. (2018). Islam and economic performance: Historical and cial development. Oxford Economic Papers, 60(4), 567–596. contemporary links. Journal of Economic Literature, 56(4), Guiso, L., Sapienza, P., & Zingales, L. (2003). People's Opium? Reli- 1292–1359. gion and economic attitudes. Journal of Monetary Economics, Lai, B., & Slater, D. (2005). Institutions of the offensive. American 50(1), 225–282. Journal of Political Science, 50(1), 113–126. Hadenius, A., & Teorell, J. (2007). Pathway from authoritarianism. Landau, D. (1986). Government and economic growth in the less Journal of Democracy, 18(1), 143–157. developed countries: An empirical study for 1960-1980. Eco- Haggard, S. (1990). Pathways from periphery. The politics of growth nomic Development and Cultural Change, 35(1), 35–75. in the newly industrializing countries. Ithaca, NY: Cornell Uni- LaPorta, R. L.-d.-S., Shleifer, A., & Vishny, R. (1997). Legal determi- versity Press. nants of external finance. Journal of Finance, 52(3), 1131–1150. ASUTAY AND MOHD SIDEK 31

Larssons, A., & Parente, S. L. (2013). Democracy as the middle gro- Rajan, R. G., & Zingales, L. (2003). The great reversals: The politics und: A unified theory of development and political regime. of financial development. Journal of Financial Economics, 69 European Economic Review, 64,35–56. (1), 5–50. Magoloni, B. (2006). Voting for autocracy. New York: Cambridge Rethel, L. (2010). Financialization and the Malaysian political econ- University Press. omy. Globalizations, 7(4), 489–506. Marsh, R. M. (1988). Sociological explanations of economic growth. Rethel, L. (2011). Whose legitimacy? Islamic finance and the global Studies in Comparative International Development, 23(4), 41–76. financial order. Review of International Political Economy, 18 Mauro, P. (1995). Corruption and growth. Quarterly Journal of Eco- (1), 75–98. nomics, 110(3), 681–712. Reuters. (2015). Rise with Turkish Islamic banks chimes with Micco, A., Panizza, U., & Yanez, M. (2007). Bank ownership and Erdogan's ideals. performance: Does politics matter? Journal of Banking & Roodman, D. (2009). Practitioners' corner. A Note on the theme of Finance, 31(1), 219–241. too many instruments. Oxford Bulletin of Economics and Statis- Mueller, D. C. (1989). Public choice II: A revised edition of public tics, 71(1), 135–158. choice. Cambridge: Cambridge University Press. Rudnyckyj, D. (2013). From wall street to halal street: Malaysia and North, D. (1990). Institutions, institutional change and economic per- the globalization of Islamic finance. The Journal of Asian Stud- formance. Cambridge: Cambridge University Press. ies, 72(4), 831–848. Nordhaus, William D (1975). The Political Business Cycle. Review of Rudnyckyj, D. (2014). Economy in practice: Islamic finance and the Economic Studies, 42(2), 169–190. problem of market reason. American Ethnologist, 41(1), Olson, M. (1993). Dictatorship, democracy & development. Ameri- 110–127. can Political Science Review, 87, 567–576. Sapienza, P. (2004). The effects of government ownership on bank Onder, Z., & Ozyildirm, S. (2013). Role of bank credit on local lending. Journal of Financial Economics, 72(2), 357–384. growth: Do politics and crisis matter? Journal of Financial Sta- Steinberg, D. A., & Malhotra, K. (2014). The effect of authoritarian bility, 9,13–25. regime type on exchange rate policy. World Politics, 66(3), Pagano, M., & Volpin, P. (2001). The political economy of finance. 491–529. Oxford Review of Economic Policy, 17(4), 502–519. Tavares, J. A., & Wacziarg, R. (2001). How democracy affects Peacock, A. T., & Wiseman, J. (1961). The growth of public expendi- growth. European Economic Review, 45, 1341–1378. ture in the United Kingdom. Princeton, NJ: Princeton University Thomson Reuters. (2019). State of the Global Islamic Economy Press. Report 2019/2020. Persson, Torsten & Tabellini, Guido (2006). Democracy and Devel- Tongur, U., Hsu, S., & Elveren, A. Y. (2015). Military expenditures opment: The Devil in the Details. American Economic Review, and political regime: Evidence from global data. Economic 96(2), 319–324. Modelling, 44,68–79. Phillips, S.; Catao, L.; Ricci, L.; Bems, R.; Das, M.; Di Giovanni J.; Wade, R. (1990). Governing the market: Economic Theory and the Unsal, D.; Castillo, M.; Lee, J.; Rodriguez, J. & Vargas, M. role of government in West Asian industrialization. Princeton, (2013). The external balance assessment (EBA) methodology. NJ: Princeton University Press. IMF Working Paper WP/13/272. Weede, E. (1983). The impact of democracy on economic growth: Pourgerami, A. (1988). The political economy of development: A Some evidence from cross-national analysis. Kyklos, 36(1), cross-national causality test of development-democracy-growth 21–39. hypothesis. Public Choice, 58(2), 123–141. Pourgerami, A. (1991). The political economy of development. An empirical investigation of the wealth theory of democracy. Jour- How to cite this article: Asutay M, Mohd nal of Theoretical Politics, 3(2), 189–211. Sidek NZ. Political economy of Islamic banking Przeworski, A (1966). Party Systems and Economics Development, PhD Dissertation, Northwest University. growth: Does political regime and institutions, Przeworski, A., & Limongi, F. (1993). Political regimes and governance and political risks matter? Int J Fin economic growth. The Journal of Economic Perspective, 7(3), Econ. 2020;1–36. https://doi.org/10.1002/ijfe.2011 51–69. 32 ASUTAY AND MOHD SIDEK

APPENDIX TABLE A1 (Continued)

Conventional TABLE A1 Samples banks Country Islamic banks banks EghtesadNovin Bank Conventional Jordan Islamic International Arab Country Islamic banks banks Bank Bahrain Venture Capital Bank Gulf International Bank Jordon Islamic Bank Al Salam Bank Bahrain National Bank of Kuwait Gulf Investment House Commercial Bank of Bahrain Kuwait Liquidity Management Ahli United Bank Kuwait Finance House Gulf Bank of Kuwait Center Bahrain ABC Bank Citi Islamic Investment Al Ahli Bank of Kuwait International Bank Kuwait Bank ABC Islamic Bank Burgan Bank National Bank of ARCAPITA Bank Kuwait Bahrain Islamic Bank Malaysia Al Rajhi Bank Malaysia CIMB Bank AlKhair Kuwait Finance House Hong Leong Malaysia Kuwait Finance House Bahrain Bank Rakyat Maybank Al Baraka Islamic Bank Bank Muamalat Public Bank Investors Bank Bank Islam Malaysia Berhad Bangladesh Social Islami Bank Oman Ahli Bank Muscat Islami Bank Bangladesh Bank of Dhofar Egypt Al Baraka Bank Egypt Bank of Muscat Faisal Islamic Bank HSBC Oman Indonesia Bank Muamalat Indonesia ArthaGraha National Bank of Bank SyariahMandiri Bank of Central Asia Oman CIMB Niaga Pakistan Bank Islami Pakistan Bank Al Habib Bank Danamon Limited Bank Internasional Meezan Islamic Bank Bank Al Faysal Indonesia Faysal Bank Bank Al Habib Bank Mayapada Metropolitan Bank Mega MCB Bank Bank Nusantara Soneri Bank Parahyangan Bank of Punjab Bank OCBC National Bank of Bank PAN Indonesia Pakistan Bank Permata Qatar Qatar Islamic Bank Arab Bank of Bank Pundi Amman Bank Sudawesi First Finance Company Commercial Bank of Qatar Iran Bank Tejarat Qatar International Doha Bank Karafarin Bank Islamic Bank Saman Bank Qatar National Bank Bank Sepah Saudi Arabia Bank Al Jazira Arab National Bank Bank Saderat Bank Al Bilad Banque Saudi Fransi Export Development Bank Al Rajhi Banking & Riyadh Bank Iran Investment ASUTAY AND MOHD SIDEK 33

TABLE A1 (Continued) TABLE A1 (Continued)

Conventional Conventional Country Islamic banks banks Country Islamic banks banks Samba Financial Emirates Islamic Bank Commercial Bank of Group Dubai Saudi British Bank Abu Dhabi Islamic Bank First Gulf Bank Saudi Hollandi Bank Gulf Finance House Saudi Investment National Bank of Bank Abu Dhabi Sudan Al Shamal Islamic Bank National Bank of Saudi Sudanese Bank Fujairah Al Salam Bank Sudan Union National Bank Sudanese Islamic Bank United Arab Bank Savings & Social Development Bank Yemen Islamic Bank of Yemen Animal Resources Bank Shamil Bank of Yeman & Bahrain Financial Investment Bank Saba Islamic Bank Blue Nile Mashreq Bank Thadamon International Islamic Sudanese French Bank Omdurman National Bank Export Development Bank Islamic Cooperative TABLE A2 Data definitions and sources of variables Development Tadamon Islamic Bank Definition and Sudan Acronym measurement Source Al Baraka Bank Sudan liquidity Deposit/total Loan Bloomberg/IRTI Bank of Khartoum size Log of total assets Bloomberg/IRTI Faisal Islamic Bank lg Loan growth Bloomberg/IRTI Turkey Bank Asya Alternatif Bank demo Institutionalized democracy Polity IV (0 to 10) Database Al Baraka Turk Akbank Participation Bank autoc Institutionalized autocracy (0 Polity IV to10) Database Kuwait Turk Participation Sekerbank Bank polity2 Subtraction of autoc score to Polity IV demo score resulting in a Database Tekstil Bankasi unified polity scale ranging Turk Ekonomi from +10 for strongly Bankasi democratic and −10 for Turkiye Garanti strongly autocratic Bankasi durable Polity IV Turkiye Is Bankasi Database Yapive Kredi xrreg The extent of Polity IV Bankasi institutionalization of Database Turkiye Sinai executive transfers Kalkinma Bankasi (1 = unregulated; 4 = regulated) UAE Sharjah Islamic Bank Abu Dhabi Commercial Bank xrcomp The competitiveness of Polity IV executive selection Database Dubai Islamic Bank Commercial Bank (1 = selection; International 4 = competitions) (Continues) (Continues) 34 ASUTAY AND MOHD SIDEK

TABLE A2 (Continued) TABLE A2 (Continued)

Definition and Definition and Acronym measurement Source Acronym measurement Source Rule of law (Percentile rank xropen The openness of executive Polity IV 0 = lowest to recruitment (1 = closed; Database 100 = highest) 4 = open)

va Voice and accountability WGI xconst Executive constraint Polity IV (Percentile rank 0 = lowest (1 = unlimited executive Database to 100 = highest) authority; 5 = substantial limitation) ten Regime durability. The WGI number of years since the parreg Regulation of participation Polity IV last regime change or (1 = unregulated; Database number of ruling years. 0 is 5 = regulated). Measures the baseline year the degree of organization and institutionalization of g_stab Government stability (0–12; International participation re-indexed to ensure Country Risk consistency of Guide (ICRG), parcomp Competitiveness of Polity IV interpretation) The PRS Group participation Database (1 = repressed; soc Socioeconomic condition (0– ICRG, PRS 5 = competitive). Measures 12; re-indexed) the degree to which int Internal conflict (0–12; re- ICRG, PRS political participation is free indexed) from government control cor2 Corruption (0–6; re-indexed) ICRG, PRS exrec Executive recruitment Polity IV dem_acc Democratic accountability (0– ICRG, PRS Database 6; re-indexed) exconst Executive constraint (1 = no Polity IV lao Law and order (0–6; re- ICRG, PRS regular limitations on Database indexed) executive actions; 5 = substation limitation) pr Political rights (0 = highest Freedom House rights; 7 = lowest rights) polcomp Political competition Polity IV (1 = uncompetitive; Database cl Civil liberties (0 = highest Freedom House 5 = competitive) rights; 7 = lowest rights) elec Year of election Database of a Electoral process (0 = no Freedom House Political election; 15 = vote freely in Institutions, legitimate elections) World Bank b Political pluralism and Freedom House (DPI) participation (0 = no b_elec Year before election DPI political pluralism; 15 = participate freely in p_elec Year after election DPI political process) cor Control of corruption Worldwide c Functioning of the Freedom House (Percentile rank 0 = lowest Governance government (0 = no to 100 = highest) Index, World representative; 15 = have Bank (WGI) representative that are g_eff Government effectiveness WGI accountable to them) (Percentile rank 0 = lowest d Freedom of expression & Freedom House to 100 = highest) belief (0 = no freedom; p_stab Political stability & absence of WGI 15 = exercise freedom of violence/terrorism expression & belief) (Percentile rank 0 = lowest e Associational and Freedom House to 100 = highest) organizational rights reg Regulatory quality (Percentile WGI (0 = not free; 15 = able to rank 0 = lowest to freely assemble and 100 = highest) associate) rol WGI ASUTAY AND MOHD SIDEK 35

TABLE A2 (Continued)

Definition and Acronym measurement Source f Rule of law (0 = no access; Freedom House 15 = access to an established and equitable system of rule of law) g Personal autonomy and Freedom House individual rights (0 = no freedom; 15 = enjoy social and economic freedoms, including equal access to economic opportunities and the right to hold private property)

TABLE A3 Correlation matrix

lg ta dta polity2 demo auto pr cl cor g_eff p_stab lg 1.0000 ta −0.0133 1.0000 dta 0.0025 −0.3899 1.0000 polity2 0.0735 −0.3045 0.0966 1.0000 demo 0.0400 0.6233 −0.3826 0.2809 1.0000 auto −0.0303 0.7558 −0.3936 −0.6340 0.5633 1.0000 pr −0.0811 −0.1501 0.2165 −0.6715 −0.7344 −0.0226 1.0000 cl −0.0776 0.0031 0.1714 −0.6195 −0.5969 0.0415 0.9045 1.0000 cor −0.0184 −0.5611 0.1539 −0.0055 −0.4539 −0.3424 0.0651 −0.1428 1.0000 g_eff −0.0025 −0.4691 0.0897 0.1764 −0.1935 −0.2913 −0.1631 −0.3778 0.8838 1.0000 p_stab −0.0153 −0.5568 0.1684 −0.0098 −0.5066 −0.3817 0.0926 −0.1237 0.9156 0.8551 1.0000 reg −0.0003 −0.5536 0.1348 0.2043 −0.2343 −0.3472 −0.1579 −0.3594 0.8607 0.8914 0.8038 rol −0.0195 −0.5841 0.1589 0.0807 −0.3687 −0.3491 −0.0302 −0.2410 0.9163 0.8992 0.8970 va 0.0650 −0.0608 −0.2073 0.6171 0.5418 −0.0815 −0.8928 −0.9145 0.2710 0.4615 0.2364 ten −0.0269 0.2295 −0.1559 −0.3386 0.2633 0.5174 −0.1015 −0.1573 0.2651 0.2791 0.2393 g_stab 0.0191 −0.2669 0.2053 −0.2483 −0.4021 −0.1050 0.3497 0.2302 0.3443 0.1959 0.4053 soci 0.0033 −0.4902 0.1656 0.0547 −0.3667 −0.3333 −0.0093 −0.2018 0.7543 0.7612 0.7821 cor2 −0.0005 −0.1094 0.0162 0.1403 0.0939 −0.0321 −0.3420 −0.4336 0.5349 0.6085 0.5118 lao 0.0412 −0.7073 0.2641 0.2135 −0.4275 −0.5107 −0.0494 −0.1788 0.6865 0.6167 0.6456 dem_acc 0.0369 −0.1049 0.1673 0.1682 0.0128 −0.1238 −0.3132 −0.3103 0.3432 0.3291 0.3259 int_con 0.0120 −0.4031 0.1602 0.0213 −0.2876 −0.2383 −0.0268 −0.1709 0.6644 0.6799 0.7715

reg rol va ten g_stab soci cor2 lao dem_acc int_con lg ta dta polity2 demo auto (Continues) 36 ASUTAY AND MOHD SIDEK

TABLE A3 (Continued)

reg rol va ten g_stab soci cor2 lao dem_acc int_con pr cl cor g_eff p_stab reg 1.0000 rol 0.9281 1.0000 va 0.4473 0.3402 1.0000 ten 0.3307 0.3449 0.1648 1.0000 g_stab 0.2654 0.2650 −0.1585 0.1939 1.0000 soci 0.7358 0.8090 0.2392 0.2532 0.2144 1.0000 cor2 0.6526 0.6211 0.4527 0.4636 0.0447 0.6278 1.0000 lao 0.7687 0.7539 0.2349 0.1960 0.2139 0.6701 0.5166 1.0000 dem_acc 0.3122 0.3514 0.3328 0.2218 0.0590 0.3660 0.4289 0.3829 1.0000 int_con 0.6340 0.7227 0.2429 0.2609 0.3924 0.7528 0.4722 0.5166 0.3673 1.0000

TABLE A4 Panel unit root diagnostics—At level TABLE A4 (Continued)

Variable tHS tDH Variable tHS tDH lg −2.312 −1.437 (0.000) (0.000) (0.000) (0.000) ten −1.222 −1.401 ta −4.2322 −2.749 (0.000) (0.000) (0.000) (0.9590) g_stab −1.668 −1.592 dtl −3.856 −3.213 (0.000) (0.000) (0.000) (0.000) socio −3.423 −3.769 polity2 −1.9093 −3.433 (0.000) (0.000) (0.0281) (0.099) int_con −2.321 −2.008 demo −2.8029 −2.543 (0.000) (0.000) (0.003) (0.000) cor2 −3.533 −2.996 auto −4.011 −3.997 (0.000) (0.000) (0.000) (0.000) dem_acc −3.713 −3.559 pr −1.949 −1.87 (0.000) (0.000) (0.005) (0.005) lao −2.573 −2.237 cl −4.737 −2.739 (0.000) (0.002) (0.000) (0.003) Note: t and t represent Herwartz and Siedenburg (2008) and − − HS DH cor 4.441 4.102 Demetrescu and Hanck (2012) tests, respectively. (0.000) (0.000) g_eff −2.398 −2.563 (0.002) (0.000) p_stab −3.997 −3.419 (0.000) (0.000) reg −1.985 −2.114 (0.000) (0.001) rol −2.111 −2.148 (0.000) (0.010) va −2.3444 −2.144